From 5f875af8a5daf962f1a13a14cc7b105265190e24 Mon Sep 17 00:00:00 2001 From: chenyan Date: Wed, 18 Nov 2026 20:17:32 +0800 Subject: [PATCH 1/2] openrouter-decisions 0.1.0: OpenRouter's Decisions API as System One A provider exposing OpenRouter's Decisions API (POST https://openrouter.ai/api/alpha/decisions) through magpie's System One interface, serving Microsoft's decision models microsoft-decision-1.1 and microsoft-decision-1 as classifiers and decision-group members. - choice, score and noul questions with probabilities, confidence and the score legend validated and restored - choices [...] shorthand expanded to a criteria map; blank criteria text falls back to the option name so the upstream's non-empty constraint holds - object, numeric and boolean state become JSON text evidence; text and multimodal part arrays pass through - per-question refusals stay refusals; other answers and usage kept - OpenRouter credit usage through auth.usage (/api/v1/auth/key, /api/v1/credits) - fixed upstream endpoint, fresh header set, no redirects; HTTP error answers never retried, one quick network-level retry for transient connection failures; cancellation propagates - decision models stay out of the chat catalog and refuse conversation 16 offline tests (node --test, bun test) with injected fetches; verified live on magpie 0.1.4462 as an intent classifier and through /provider test (decide 2228 ms) in a sandbox HOME. --- README.md | 1 + packages/openrouter-decisions/CHANGELOG.md | 20 + packages/openrouter-decisions/LICENSE | 21 + packages/openrouter-decisions/NOTICE.md | 9 + packages/openrouter-decisions/README.md | 200 +++ packages/openrouter-decisions/VERIFICATION.md | 33 + packages/openrouter-decisions/adapter.mjs | 231 +++ packages/openrouter-decisions/index.mjs | 103 ++ packages/openrouter-decisions/opencode.mjs | 1 + packages/openrouter-decisions/opencode2.mjs | 620 +++++++ packages/openrouter-decisions/package.json | 19 + .../test/adapter.test.mjs | 186 ++ registry.json | 1597 ++++++++++++++--- 13 files changed, 2811 insertions(+), 230 deletions(-) create mode 100644 packages/openrouter-decisions/CHANGELOG.md create mode 100644 packages/openrouter-decisions/LICENSE create mode 100644 packages/openrouter-decisions/NOTICE.md create mode 100644 packages/openrouter-decisions/README.md create mode 100644 packages/openrouter-decisions/VERIFICATION.md create mode 100644 packages/openrouter-decisions/adapter.mjs create mode 100644 packages/openrouter-decisions/index.mjs create mode 100644 packages/openrouter-decisions/opencode.mjs create mode 100644 packages/openrouter-decisions/opencode2.mjs create mode 100644 packages/openrouter-decisions/package.json create mode 100644 packages/openrouter-decisions/test/adapter.test.mjs diff --git a/README.md b/README.md index 42c5bd65c..f0f88584c 100644 --- a/README.md +++ b/README.md @@ -40,6 +40,7 @@ vendor service and makes its requests. The packages work in OpenCode and in | [mimo](packages/mimo) | Xiaomi MiMo, with a Xiaomi account | `mimo-app` | | [nous](packages/nous) | Nous Portal models, with Nous Portal's device sign-in | `nous-portal` | | [openai-decisions](packages/openai-decisions) | OpenAI's native Decisions API as System One for magpie's classifiers and decision groups, with standard API-key sign-in and per-question refusal handling | `openai-decisions` | +| [openrouter-decisions](packages/openrouter-decisions) | OpenRouter's Decisions API (Microsoft Decision 1.1 and 1) as System One for magpie's classifiers and decision groups, with API-key sign-in, credit usage and per-question refusal handling | `openrouter-decisions` | | [opendesign](packages/opendesign) | Open Design Cloud (open-design.ai), with the Open Design CLI's device sign-in | `opendesign` | | [qoder](packages/qoder) | Qoder subscription (qoder.com), with Qoder's device sign-in | `qoder` | | [qoder](packages/qoder) | Qoder CN subscription (qoder.cn), with Qoder CN's device sign-in | `qoder-cn` | diff --git a/packages/openrouter-decisions/CHANGELOG.md b/packages/openrouter-decisions/CHANGELOG.md new file mode 100644 index 000000000..70c8f6a90 --- /dev/null +++ b/packages/openrouter-decisions/CHANGELOG.md @@ -0,0 +1,20 @@ +# Changelog + +## 0.1.0 — 2026-11-18 + +First release. Exposes OpenRouter's Decisions API +(`POST https://openrouter.ai/api/alpha/decisions`) through Magpie's System +One interface, with `microsoft/microsoft-decision-1.1` as the model. + +- `choice`, `score` and `noul` questions with probabilities, confidence and + a score legend restored from the upstream answer. +- `choices: [...]` shorthand expanded to a criteria map; blank criteria text + falls back to the option name so the upstream's non-empty constraint holds. +- Object, numeric and boolean System One state becomes JSON text evidence; + text passes through; multimodal state parts keep their array form. +- Per-question refusals stay refusals; other answers and usage are retained. +- OpenRouter credit usage shown through `auth.usage` (`/api/v1/auth/key`, + `/api/v1/credits`). +- Fixed upstream endpoint, fresh header set, no redirects; cancellation + propagates; one quick network-level retry absorbs momentary proxy or DNS + failures while HTTP error answers are never retried. diff --git a/packages/openrouter-decisions/LICENSE b/packages/openrouter-decisions/LICENSE new file mode 100644 index 000000000..e6bc674fe --- /dev/null +++ b/packages/openrouter-decisions/LICENSE @@ -0,0 +1,21 @@ +MIT License + +Copyright (c) 2026 magpie-community contributors + +Permission is hereby granted, free of charge, to any person obtaining a copy +of this software and associated documentation files (the "Software"), to deal +in the Software without restriction, including without limitation the rights +to use, copy, modify, merge, publish, distribute, sublicense, and/or sell +copies of the Software, and to permit persons to whom the Software is +furnished to do so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. diff --git a/packages/openrouter-decisions/NOTICE.md b/packages/openrouter-decisions/NOTICE.md new file mode 100644 index 000000000..c6cd4f06a --- /dev/null +++ b/packages/openrouter-decisions/NOTICE.md @@ -0,0 +1,9 @@ +The OpenCode 2.x compatibility bridge, `opencode2.mjs`, is an unmodified +copy of `scripts/opencode2.mjs` from magpie-community/plugins, distributed +under that repository's MIT license. Keep it byte-identical to that copy. + +Upstream: https://github.com/magpie-community/plugins + +The implementation follows magpie's public provider/auth/decision contracts +and OpenRouter's public Decisions API. No third-party provider application +logic is bundled here. diff --git a/packages/openrouter-decisions/README.md b/packages/openrouter-decisions/README.md new file mode 100644 index 000000000..b6e6820d8 --- /dev/null +++ b/packages/openrouter-decisions/README.md @@ -0,0 +1,200 @@ +# OpenRouter Decisions for Magpie + +A provider exposing **OpenRouter's [Decisions API](https://openrouter.ai/docs/guides/features/decisions)** +through Magpie's **System One** interface. Use it as a chat routing group's +classifier (intent or reasoning effort) or as a member of a decision-model +group with other System One providers. + +**Provider id:** `openrouter-decisions`. **Models:** +`openrouter-decisions/microsoft-decision-1.1` and +`openrouter-decisions/microsoft-decision-1` (Microsoft's System One decision +models on OpenRouter; 1.1 reads text and images, 1 is text-only and cheaper). + +Magpie already speaks System One for compatible decision providers. OpenRouter +speaks the same question/answer dialect on a different path with a few extra +constraints, which this plugin converts: + +| System One | OpenRouter Decisions | +| --- | --- | +| `POST /v1/systemone` (via the plugin's `fetch`) | `POST https://openrouter.ai/api/alpha/decisions` | +| `model: "microsoft-decision-1.1"` | `model: "microsoft/microsoft-decision-1.1"` | +| object / number / boolean `state` | JSON text `state` | +| `choice` with a `choices` array | expanded to a `criteria` map | +| blank criteria text | the option name, then a positional label | +| `score` with ordered criteria | same array, blank levels labeled | +| `noul` with optional true/false criteria | same fields | + +Answers are converted back with validation: probabilities must cover every +option and sum to one (±0.02), a `score` must land inside `0..N-1`, and a +per-question `refusal` stays a refusal while other answers and usage are +kept. + +## Install and sign in + +Install a released community package through Magpie: + +```sh +magpie plugin add @magpie-community/opencode-openrouter-decisions-auth +magpie plugin login openrouter-decisions +``` + +For an unpublished package or a source checkout, run from the root of +`magpie-community/plugins` instead: + +```sh +magpie plugin add ./packages/openrouter-decisions +magpie plugin login openrouter-decisions +``` + +In Magpie's Plugins page, install **OpenRouter Decisions**, then sign in with +an OpenRouter API key (`sk-or-…`). The key is kept in Magpie's +`plugin-auth.json` (Unix permissions 0600); requests use the selected key's +OpenRouter allowance. + +## Use as a classifier + +As an intent classifier: + +```sh +magpie group add my-group models=provider/model-a,provider/model-b routing=order +magpie group rule add my-group use=provider/model-a intent="a hard task: designing or debugging" +magpie group set my-group classifier=openrouter-decisions/microsoft-decision-1.1 +``` + +As an effort classifier (the group needs at least two models, and members +need reasoning levels): + +```sh +magpie group set my-group classifier=openrouter-decisions/microsoft-decision-1.1 effort=auto +``` + +## Call System One directly + +```sh +curl -s -X POST http://127.0.0.1:3425/v1/systemone \ + -H "Authorization: Bearer $MAGPIE_KEY" -H "Content-Type: application/json" \ + -d '{ + "model": "openrouter-decisions/microsoft-decision-1.1", + "state": {"message": "What does HTTP 404 mean?"}, + "questions": { + "is_question": {"type": "noul", "instructions": "Does this ask a question?"}, + "intent": { + "type": "choice", + "instructions": "Which kind of request is this?", + "criteria": {"question": "Asks for an explanation", "code_change": "Asks to modify code"} + }, + "effort": { + "type": "score", + "instructions": "How much work does this request require?", + "criteria": ["Little work", "Some work", "Much work", "Most work"] + } + } + }' +``` + +A real response: + +```json +{ + "model": "microsoft/microsoft-decision-1.1-20261117", + "answers": { + "is_question": {"type": "noul", "noul": 0.998}, + "intent": {"type": "choice", "choice": "question", + "probabilities": {"question": 0.998, "code_change": 0.0017}, "confidence": 0.9966}, + "effort": {"type": "score", "score": 0.179, + "legend": {"0": "Little work", "1": "Some work", "2": "Much work", "3": "Most work"}, + "probabilities": {"0": 0.858, "1": 0.112, "2": 0.021, "3": 0.0086}, "confidence": 0.811} + }, + "usage": {"input_tokens": 114, "output_tokens": 3, "cost": 0.000004788}, + "provider": "Azure" +} +``` + +## Usage + +`auth.usage` reads OpenRouter's `/api/v1/auth/key` and `/api/v1/credits`; +`magpie quota` shows the credit window, its percentage and the remaining +balance. + +## Protocol and limits + +- Only `POST /v1/systemone` is accepted; other paths and methods are refused + with 400 and never sent upstream. +- The upstream request goes to the fixed `https://openrouter.ai/api/alpha/decisions` + with a fresh header set, no redirects, and no gateway headers or cookies. +- 1–400 questions, 2–255 choice options, 2–10 ordered score levels. +- Upstream error status, body, `Retry-After`, `x-request-id` and + `x-ratelimit-*` headers pass through unchanged. +- Cancellation propagates; HTTP error answers are never retried. One quick + network-level retry absorbs a momentary proxy or DNS failure. +- The decision models stay out of the chat catalog and refuse conversation. +- The adapter does not retry rate limits and owns no failover; Magpie routes. + +## Develop and contribute + +```sh +npm test # or: node --test test/*.test.mjs +bun scripts/check.mjs openrouter-decisions # from the community repo root +MAGPIE=/absolute/path/to/magpie scripts/try.sh openrouter-decisions +``` + +The tests are offline: credentials and responses are synthetic and in-memory, +and no test writes a credential file or touches the network. + +[index.mjs](index.mjs) registers the provider, the selected-account auth and +the usage hook. [adapter.mjs](adapter.mjs) converts requests and answers and +sends the fixed upstream request. Keep `opencode2.mjs` byte-identical to the +community's canonical `scripts/opencode2.mjs`. See [NOTICE.md](NOTICE.md) +for attribution and [VERIFICATION.md](VERIFICATION.md) for live evidence and +untested scope. + +Review checks: + +- Receive credentials only through the host's selected-account auth. +- Preserve probabilities, confidence, the score legend and per-question + refusals. +- Supply non-empty instructions and criteria wherever the upstream requires + them. +- Keep tests offline and free of credential files; code comments use English. +- Keep the root README, package README, bilingual registry and changelog + aligned. +- Verify live routing separately from fixtures; pace live checks by actual + quota. +- Keep credentials, internal endpoints and local artifacts outside the + package. + +References: [OpenRouter Decisions](https://openrouter.ai/docs/guides/features/decisions), +[Magpie provider contract](https://usemagpie.ai/docs/plugins#provider). + +MIT. Not affiliated with OpenRouter or Microsoft. + +## 中文 + +安装社区发行包并使用 OpenRouter API key 登录: + +```sh +magpie plugin add @magpie-community/opencode-openrouter-decisions-auth +magpie plugin login openrouter-decisions +``` + +未发布时,在社区仓库根目录使用 +`magpie plugin add ./packages/openrouter-decisions`,再执行同一登录命令。 +提供方 ID 为 `openrouter-decisions`,模型为 +`openrouter-decisions/microsoft-decision-1.1` 与 +`openrouter-decisions/microsoft-decision-1`。 + +统一入口为 Magpie `/v1/systemone`。支持选择、评分和是非概率(含概率分布、 +置信度与评分档位说明);单题拒绝(refusal)原样保留,其余答案与用量不受影响。 +作为意图分类器时,路由记录会显示命中的意图与 `sure` 置信度;作为推理强度 +分类器时,用评分题给每一轮打难度(组内需至少两个带推理档位的模型)。 + +凭据由宿主提供当前选中账号的 API key;用量钩子读取 OpenRouter 的额度与 +余额,显示在 `magpie quota` 与供应商卡片上。只接受发往固定上游地址的请求, +错误状态码、`Retry-After` 与限流头原样透传;瞬时网络故障自动快速重试一次, +HTTP 错误回答不重试,故障切换交给 magpie。 + +## 客户端兼容性 + +- Magpie:完整支持(意图分类器、推理强度分类器、决策模型组、`/v1/systemone`)。 +- OpenCode 1.x:加载与登录可用;决策输出需要客户端自己发起决策调用。 +- OpenCode 2.x:通过 `opencode2.mjs` 桥接加载;决策工作流未端到端验证。 diff --git a/packages/openrouter-decisions/VERIFICATION.md b/packages/openrouter-decisions/VERIFICATION.md new file mode 100644 index 000000000..303f42791 --- /dev/null +++ b/packages/openrouter-decisions/VERIFICATION.md @@ -0,0 +1,33 @@ +# Verification + +## Initial release — 2026-11-18 + +The adapter's 15 offline tests pass on Node (`node --test test/*.test.mjs`). +They cover the request mapping (state forms, the `choices` shorthand, blank +criteria fallbacks, instruction defaults), answer validation (probability +distributions, score ranges, refusals, malformed answers), endpoint and +header hygiene (fixed upstream URL, fresh header set, no forwarded gateway +headers or cookies), upstream error passthrough, credential loss, the +network retry, and cancellation. + +Live end-to-end checks ran on magpie 0.1.4462 with a real OpenRouter key: + +- `POST /v1/systemone` with `choice`, `score` and `noul` questions returned + calibrated probabilities, the score legend and usage (`provider: "Azure"`). +- As an intent classifier (`magpie group rule add … intent=…` with + `classifier=openrouter-decisions/microsoft-decision-1.1`), the decision + reached the gateway in ~1.1–3.0 s and the recorded `classified` entry + carried the matched intent and `sure` values of 0.94 and 0.99. +- As an effort classifier (`magpie group set … effort=auto`), the + group saved and the provider registered with `decides: true`; the model + stays out of `/v1/models` and refuses conversation requests. +- `magpie quota` showed the OpenRouter credit window and balance through + the usage hook. + +A transient network failure against the proxy was observed in production +use (`Could not reach OpenRouter` after ~3 s); the next call succeeded. The +single quick retry was added for that case and is covered by a test. + +Not verified: OpenCode 2.x decision workflows end to end, other OpenRouter +decision models, image (`box`) inputs — the model does not declare them and +the adapter does not accept `box` questions. diff --git a/packages/openrouter-decisions/adapter.mjs b/packages/openrouter-decisions/adapter.mjs new file mode 100644 index 000000000..34687da5a --- /dev/null +++ b/packages/openrouter-decisions/adapter.mjs @@ -0,0 +1,231 @@ +// Converts Magpie's System One requests into OpenRouter's Decisions API +// (POST https://openrouter.ai/api/alpha/decisions) and converts the answers back. +// OpenRouter already speaks a System One dialect, so the mapping is shallow: +// validate, translate the state into a supported form, and verify answers. + +// The models this provider serves. Both are Microsoft's System One decision +// models on OpenRouter; 1.1 reads text and images, 1 is text-only and cheaper. +export const MODELS = { + "microsoft-decision-1.1": { upstream: "microsoft/microsoft-decision-1.1", images: true }, + "microsoft-decision-1": { upstream: "microsoft/microsoft-decision-1", images: false }, +}; +export const MODEL = "microsoft-decision-1.1"; +const BASE = "https://openrouter.ai/api/v1"; +const SYSTEM_ONE_URL = `${BASE}/systemone`; +const DECISIONS_URL = "https://openrouter.ai/api/alpha/decisions"; +const FALLBACK_INSTRUCTIONS = "Answer this question using the supplied input and criteria."; + +const record = (value) => value !== null && typeof value === "object" && !Array.isArray(value); + +function requireValue(condition, message) { + if (!condition) throw new Error(message); +} + +function fields(value, allowed, name) { + requireValue(record(value), `${name} must be an object`); + requireValue(Object.keys(value).every((key) => allowed.includes(key)), `${name} contains an unsupported field`); +} + +function description(value) { + return typeof value === "string" ? value : JSON.stringify(value); +} + +function questionInstructions(question) { + const supplied = question.instructions == null ? "" : description(question.instructions); + return supplied.trim() ? supplied : FALLBACK_INSTRUCTIONS; +} + +// Magpie allows `choices: ["a", "b"]` as a shorthand for a criteria map. +// OpenRouter's Decisions API documents the criteria form only, so expand it. +function choiceCriteria(question) { + if (question.choices !== undefined) { + requireValue(question.criteria === undefined, "choice accepts criteria or choices, not both"); + requireValue(Array.isArray(question.choices) && question.choices.every((value) => typeof value === "string"), "choices must be strings"); + requireValue(question.choices.every((value) => value.length > 0), "Choice names must not be empty"); + requireValue(new Set(question.choices).size === question.choices.length, "choices must be unique"); + return Object.fromEntries(question.choices.map((value) => [value, null])); + } + requireValue(record(question.criteria), "choice requires a criteria object"); + return question.criteria; +} + +// OpenRouter rejects blank criteria text ("String should have at least 1 +// character"), while System One allows a bare option name or a null level. +// Fall back to the option name, then to a positional label. +function criterionText(value, fallback) { + const text = value == null ? "" : description(value); + return text.trim() ? text : fallback; +} + +export function toOpenRouterRequest(body) { + fields(body, ["model", "state", "questions"], "request"); + const spec = MODELS[body.model]; + requireValue(spec, `Only ${Object.keys(MODELS).join(" and ")} are supported`); + requireValue(Object.hasOwn(body, "state") && body.state !== null, "state is required"); + requireValue(record(body.questions), "questions must be an object"); + const entries = Object.entries(body.questions); + requireValue(entries.length > 0 && entries.length <= 400, "Expected 1–400 questions"); + const questions = {}; + for (const [name, question] of entries) { + requireValue(name.length > 0, "Question IDs must not be empty"); + fields(question, ["type", "instructions", "criteria", "choices"], "question"); + requireValue(question.instructions == null || typeof question.instructions === "string" || typeof question.instructions === "object", "instructions must be text or structured JSON"); + const common = { type: question.type, instructions: questionInstructions(question) }; + switch (question.type) { + case "choice": { + const criteria = choiceCriteria(question); + const keys = Object.keys(criteria); + requireValue(keys.length >= 2 && keys.length <= 255, "choice requires 2–255 options"); + requireValue(keys.every((key) => key.length > 0), "Choice names must not be empty"); + requireValue(new Set(keys).size === keys.length, "Choice names must be unique"); + questions[name] = { ...common, criteria: Object.fromEntries(keys.map((key) => [key, criterionText(criteria[key], key)])) }; + break; + } + case "score": + requireValue(question.choices === undefined, "score does not accept choices"); + requireValue(Array.isArray(question.criteria) && question.criteria.length >= 2 && question.criteria.length <= 10, "score requires 2–10 ordered criteria"); + questions[name] = { ...common, criteria: question.criteria.map((value, index) => criterionText(value, `Level ${index}`)) }; + break; + case "noul": + requireValue(question.choices === undefined, "noul does not accept choices"); + if (question.criteria !== undefined) { + requireValue(record(question.criteria) && Object.keys(question.criteria).every((key) => key === "true" || key === "false"), "noul criteria accept true and false only"); + questions[name] = { + ...common, + criteria: Object.fromEntries(Object.entries(question.criteria).map(([key, value]) => [key, criterionText(value, key === "true" ? "True" : "False")])), + }; + } else { + questions[name] = common; + } + break; + default: + throw new Error("Unsupported question type; use choice, score or noul"); + } + } + // OpenRouter accepts a text state, or an array of text/media parts. Anything + // else (plain objects, numbers, booleans) becomes JSON text evidence. The + // text-only model takes every part as its JSON text. + const state = typeof body.state === "string" + ? body.state + : (Array.isArray(body.state) ? body.state.map((part) => typeof part === "string" ? part : JSON.stringify(part)) : JSON.stringify(body.state)); + return { model: spec.upstream, state, questions }; +} + +function probability(value, name) { + requireValue(typeof value === "number" && Number.isFinite(value) && value >= 0 && value <= 1, `Invalid ${name}`); + return value; +} + +function distribution(map, expected, name) { + requireValue(record(map), `${name} probabilities must be an object`); + const keys = Object.keys(map); + requireValue(keys.length === expected.length && expected.every((key) => Object.hasOwn(map, key)), `Incomplete ${name} probability distribution`); + let sum = 0; + const result = {}; + for (const key of expected) { + result[key] = probability(map[key], name); + sum += result[key]; + } + requireValue(Math.abs(sum - 1) <= 0.02, `${name} probabilities do not sum to one`); + return result; +} + +export function fromOpenRouterResponse(upstream, source) { + const request = toOpenRouterRequest(source); + requireValue(record(upstream), "Invalid upstream decision response"); + requireValue(record(upstream.answers), "Invalid upstream answers"); + const answers = {}; + for (const [name, question] of Object.entries(request.questions)) { + const answer = upstream.answers[name]; + requireValue(record(answer), `Missing upstream answer for "${name}"`); + if (answer.type === "refusal") { + answers[name] = { type: "refusal" }; + continue; + } + requireValue(answer.type === question.type, `Wrong upstream answer type for "${name}"`); + if (answer.type === "noul") { + answers[name] = { type: "noul", noul: probability(answer.noul, "noul") }; + } else if (answer.type === "choice") { + const keys = Object.keys(request.questions[name].criteria); + requireValue(keys.includes(answer.choice), "Upstream chose an unknown option"); + answers[name] = { + type: "choice", + choice: answer.choice, + probabilities: distribution(answer.probabilities, keys, "choice"), + confidence: probability(answer.confidence, "confidence"), + }; + } else { + const count = request.questions[name].criteria.length; + requireValue(typeof answer.score === "number" && Number.isFinite(answer.score) && answer.score >= 0 && answer.score <= count - 1, "Invalid upstream score"); + answers[name] = { + type: "score", + score: answer.score, + legend: Object.fromEntries(request.questions[name].criteria.map((text, index) => [String(index), text])), + probabilities: distribution(answer.probabilities, request.questions[name].criteria.map((_, index) => String(index)), "score"), + confidence: probability(answer.confidence, "confidence"), + }; + } + } + return { ...upstream, answers }; +} + +function responseHeaders(response) { + const headers = new Headers({ "content-type": response.headers.get("content-type") || "application/json" }); + for (const [key, value] of response.headers) { + if (key === "retry-after" || key === "x-request-id" || key.startsWith("x-ratelimit-")) headers.set(key, value); + } + return headers; +} + +function refused(status, message, headers) { + return Response.json({ error: { type: status === 400 ? "invalid_request_error" : "api_error", message } }, { status, headers }); +} + +export function createDecisionFetch({ getKey, fetchImpl = globalThis.fetch }) { + return async (input, init) => { + let request, source, payload; + try { + request = new Request(input, init); + requireValue(request.url === SYSTEM_ONE_URL && request.method === "POST", "Only POST /v1/systemone is supported"); + source = await request.json(); + payload = toOpenRouterRequest(source); + } catch (error) { + if (request?.signal.aborted) throw error; + return refused(400, error instanceof SyntaxError ? "Invalid JSON request" : error.message); + } + request.signal.throwIfAborted(); + let key; + try { + key = await getKey(); + requireValue(typeof key === "string" && key.trim().length > 0 && !/[\r\n]/.test(key), "Missing OpenRouter API key"); + } catch { + return refused(503, "OpenRouter credentials unavailable; check this provider's API-key sign-in"); + } + const headers = new Headers({ "content-type": "application/json", accept: "application/json", authorization: `Bearer ${key}` }); + // A momentary network failure (proxy restart, DNS hiccup) should not lose + // the routing decision: retry network-level errors once, quickly. HTTP + // error responses are upstream answers and are never retried. + let upstream; + for (let attempt = 0; ; attempt++) { + try { + upstream = await fetchImpl(DECISIONS_URL, { method: "POST", headers, body: JSON.stringify(payload), signal: request.signal, redirect: "error" }); + break; + } catch (error) { + if (request.signal.aborted) throw error; + if (attempt > 0) return refused(502, `Could not reach OpenRouter: ${error?.message || "network error"}`); + await new Promise((resolve) => setTimeout(resolve, 400)); + } + } + const keptHeaders = responseHeaders(upstream); + if (!upstream.ok) return new Response(upstream.body, { status: upstream.status, headers: keptHeaders }); + try { + const result = fromOpenRouterResponse(await upstream.json(), source); + return Response.json(result, { status: upstream.status, headers: keptHeaders }); + } catch (error) { + if (request.signal.aborted) throw error; + return refused(502, "OpenRouter returned an invalid or incomplete decision answer", keptHeaders); + } + }; +} + +export const _internal = { SYSTEM_ONE_URL, DECISIONS_URL }; diff --git a/packages/openrouter-decisions/index.mjs b/packages/openrouter-decisions/index.mjs new file mode 100644 index 000000000..6b6465bb9 --- /dev/null +++ b/packages/openrouter-decisions/index.mjs @@ -0,0 +1,103 @@ +import { createDecisionFetch, MODELS } from "./adapter.mjs"; + +const ID = "openrouter-decisions"; +const BASE = "https://openrouter.ai/api/v1"; +// $ per 1M tokens, billed on input only; both models answer with decisions, +// never conversation. 1.1 reads text and images, 1 is text-only. +const modelSpecs = { + "microsoft-decision-1.1": { + name: "Microsoft Decision 1.1", + modalities: { input: ["text"], output: ["text"] }, + }, + "microsoft-decision-1": { + name: "Microsoft Decision 1", + modalities: { input: ["text"], output: ["text"] }, + }, +}; +const models = Object.fromEntries(Object.entries(MODELS).map(([id, spec]) => [ + id, + { + id, + name: modelSpecs[id].name, + decides: true, + tool_call: false, + reasoning: false, + limit: { context: 32768, output: 0 }, + modalities: modelSpecs[id].modalities, + cost: { input: 0.042, output: 0 }, + }, +])); + +// Export only the plugin function; Magpie invokes each exported function as a plugin. +export const OpenRouterDecisionsPlugin = async () => ({ + async config(config) { + config.provider ??= {}; + config.provider[ID] ??= { + name: "OpenRouter Decisions", + npm: "@ai-sdk/openai-compatible", + api: BASE, + decide: true, + models, + }; + }, + auth: { + provider: ID, + methods: [ + { type: "api", label: "OpenRouter API key", placeholder: "sk-or-…" }, + ], + async loader(getAuth) { + const auth = await getAuth(); + if (auth?.type !== "api" || typeof auth.key !== "string" || !auth.key.trim()) return {}; + return { + baseURL: BASE, + apiKey: auth.key, + fetch: createDecisionFetch({ getKey: async () => { + const current = await getAuth(); + return current?.type === "api" ? current.key : undefined; + } }), + }; + }, + // magpie's own hook: show this account's OpenRouter credit usage + async usage(getAuth) { + const auth = await getAuth(); + if (auth?.type !== "api" || typeof auth.key !== "string" || !auth.key.trim()) return { error: "Not signed in" }; + const headers = { authorization: `Bearer ${auth.key}` }; + try { + const res = await fetch(`${BASE}/auth/key`, { headers }); + if (!res.ok) return { error: `OpenRouter answered ${res.status}` }; + const { data } = await res.json(); + if (!data) return { error: "OpenRouter returned no key info" }; + const windows = []; + if (typeof data.limit === "number" && data.limit > 0 && typeof data.usage === "number") { + windows.push({ + name: "Credits", + used: Math.round((data.usage / data.limit) * 1000) / 10, + amount: data.usage, + limit: data.limit, + unit: "credits", + display: `${data.usage.toFixed(2)} / ${data.limit.toFixed(2)} credits`, + }); + } + const result = { plan: data.is_free_tier ? "Free tier" : "Paid", user: data.label, windows }; + try { + const creditsRes = await fetch(`${BASE}/credits`, { headers }); + if (creditsRes.ok) { + const credits = (await creditsRes.json()).data; + if (credits && typeof credits.total_credits === "number" && typeof credits.total_usage === "number") { + result.balance = `$${(credits.total_credits - credits.total_usage).toFixed(2)}`; + } + } + } catch { /* balance is best-effort */ } + return result; + } catch { + return { error: "Could not reach OpenRouter" }; + } + }, + }, + provider: { + id: ID, + async models() { + return models; + }, + }, +}); diff --git a/packages/openrouter-decisions/opencode.mjs b/packages/openrouter-decisions/opencode.mjs new file mode 100644 index 000000000..5a9a55f32 --- /dev/null +++ b/packages/openrouter-decisions/opencode.mjs @@ -0,0 +1 @@ +export { OpenRouterDecisionsPlugin } from "./index.mjs"; diff --git a/packages/openrouter-decisions/opencode2.mjs b/packages/openrouter-decisions/opencode2.mjs new file mode 100644 index 000000000..b9393085e --- /dev/null +++ b/packages/openrouter-decisions/opencode2.mjs @@ -0,0 +1,620 @@ +// The OpenCode 2.x face of a package whose plugins are written for +// OpenCode 1.x (and magpie's plugin host). +// +// OpenCode 2.x loads a package's `exports["."]` and wants its default to be +// {id, setup(ctx)}; it has no fallback for a 1.x plugin's named functions +// (plugins#657). This file is that default. It runs each 1.x plugin the +// package's opencode.mjs exports and hands 2.x what the plugin's hooks say: +// +// auth.methods → the integration's sign-in methods +// config() → the provider, its package, base URL and models +// provider.models → the signed-in account's own model list +// auth.loader → each request's base URL, key, headers and fetch +// +// What 2.x has no place for is left out: chat.headers and chat.params +// (2.x's sdk hook doesn't carry the session), magpie's usage and checkin, +// tools and events. +// +// scripts/opencode2.mjs is the one copy that is edited. Every package that +// has it carries the same bytes as packages//opencode2.mjs, and +// scripts/check.mjs fails when a copy differs. + +import { readFileSync } from "node:fs" +import * as v1 from "./opencode.mjs" + +const pkg = JSON.parse(readFileSync(new URL("./package.json", import.meta.url), "utf8")) + +// The key the SDK is made with: the account's own goes on each request in +// its place (see wrapFetch), so a key read when the SDK was made can't go +// out after the account changed. +const PLACEHOLDER = "opencode2-v1-key" + +// The SDK package a 1.x npm name runs as on 2.x. A version keeps 2.x from +// swapping @ai-sdk/* for its native providers, which never call the sdk +// hook below; openai-compatible's is the one 2.x ships, the others are AI +// SDK v6's dist-tag, the line 2.x's bridge runs. +export function packageOf(npm) { + const name = npm || "@ai-sdk/openai-compatible" + if (name === "@ai-sdk/openai-compatible") return "aisdk:@ai-sdk/openai-compatible@2.0.41" + if (/^@ai-sdk\/[^/@]+$/.test(name)) return `aisdk:${name}@ai-v6` + return `aisdk:${name}` +} + +// The plugins a 1.x module exports, as magpie's host and OpenCode 1.x read +// them: a default {server} is the one plugin, else every function export +// and every {server} object. +export function pluginsOf(mod) { + const d = mod?.default + if (d && typeof d === "object" && ("server" in d || "id" in d || "tui" in d)) + return typeof d.server === "function" ? [d.server] : [] + const out = [] + for (const value of Object.values(mod ?? {})) { + const fn = typeof value === "function" ? value : value && typeof value === "object" && typeof value.server === "function" ? value.server : null + if (fn && !out.includes(fn)) out.push(fn) + } + return out +} + +const int = (n, fallback = 0) => { + const v = Math.trunc(Number(n)) + return Number.isFinite(v) && v >= 0 ? v : fallback +} +const money = (n) => (Number.isFinite(Number(n)) && Number(n) >= 0 ? Number(n) : 0) +const record = (v) => (v && typeof v === "object" && !Array.isArray(v) ? v : {}) +const json = (v) => { + try { + return JSON.parse(JSON.stringify(v ?? null)) + } catch { + return null + } +} +const strings = (v) => Object.fromEntries(Object.entries(record(v)).filter(([, s]) => typeof s === "string")) +const nonEmpty = (o) => (Object.keys(o).length ? o : undefined) + +// connKey names a connection the way 2.x does. +const connKey = (c) => (!c ? "" : c.type === "credential" ? `credential:${c.id}` : `env:${c.name}`) + +// ---- sign-in ------------------------------------------------------------ + +// fields is a 1.x method's prompts as a 2.x form. +export function fieldsOf(prompts) { + const out = [] + for (const p of prompts ?? []) { + if (!p?.key) continue + const f = { key: String(p.key), type: "string", title: String(p.message ?? p.key) } + if (p.placeholder) f.placeholder = String(p.placeholder) + if (p.type === "select" && Array.isArray(p.options)) + f.options = p.options.map((o) => ({ value: String(o.value), label: String(o.label ?? o.value), ...(o.hint ? { description: String(o.hint) } : {}) })) + // a prompt shown only when another answer says so can't be required, and + // neither can one whose own validate() accepts an empty answer (the + // browser sign-in's optional API key) + if (!p.condition && p.required !== false) f.required = true + out.push(f) + } + return out.length ? out : undefined +} + +// v1Saved is what a 1.x sign-in's success is saved as, the way 1.x writes +// auth.json: the result without its type and provider. +function v1Saved(r) { + const { type: _t, provider: _p, ...rest } = r + if ("key" in r) return { type: "api", ...rest } + return { type: "oauth", ...rest } +} + +// credentialOf is a 1.x sign-in's success as a 2.x credential: OAuth for +// both kinds, the 1.x auth kept whole in its metadata. No refresh is +// registered, so 2.x hands the credential over as it is and the plugin's +// own loader renews it, as it does on 1.x. +export function credentialOf(methodID, r) { + if (!r || r.type !== "success") throw new Error(r?.error || r?.message || "the sign-in failed") + const saved = json(v1Saved(r)) + if (saved.type === "api") + return { type: "oauth", methodID, access: String(saved.key ?? ""), refresh: "", expires: 0, metadata: { v1: saved } } + return { + type: "oauth", + methodID, + access: String(saved.access ?? ""), + refresh: String(saved.refresh ?? ""), + expires: int(saved.expires), + metadata: { v1: saved }, + } +} + +// authOfValue is the 1.x auth a 2.x credential stands for. +export function authOfValue(value) { + if (!value) return null + if (value.type === "oauth") { + const v1 = record(value.metadata).v1 + if (v1 && typeof v1 === "object") return v1 + return { type: "oauth", access: value.access, refresh: value.refresh, expires: value.expires } + } + if (value.type === "key") { + const meta = value.metadata && Object.keys(value.metadata).length ? { metadata: value.metadata } : {} + return { type: "api", key: value.key, ...meta } + } + return null +} + +const fingerprint = (a) => JSON.stringify([a?.type, a?.access ?? a?.key ?? "", a?.refresh ?? ""]) + +// ---- models ------------------------------------------------------------- + +const MODALITIES = ["text", "audio", "image", "video", "pdf"] + +// internalOf is a config model as 1.x hands it to provider.models: what +// fromModelsDevModel makes of it, with the config's own variants. +export function internalOf(providerID, conf, key, m) { + const mods = record(m.modalities) + const has = (list, k) => Array.isArray(list) && list.includes(k) + const cost = record(m.cost) + return { + ...m, + id: key, + providerID, + name: m.name ?? key, + api: { id: m.id ?? key, url: m.provider?.api ?? conf.api ?? conf.options?.baseURL ?? "", npm: m.provider?.npm ?? conf.npm ?? "@ai-sdk/openai-compatible" }, + status: m.status ?? "active", + headers: m.headers ?? {}, + options: m.options ?? {}, + cost: { input: money(cost.input), output: money(cost.output), cache: { read: money(cost.cache_read), write: money(cost.cache_write) } }, + limit: { context: int(m.limit?.context), input: m.limit?.input, output: int(m.limit?.output) }, + capabilities: { + temperature: m.temperature ?? false, + reasoning: m.reasoning ?? false, + attachment: m.attachment ?? false, + toolcall: m.tool_call ?? true, + input: Object.fromEntries(MODALITIES.map((k) => [k, mods.input ? has(mods.input, k) : k === "text"])), + output: Object.fromEntries(MODALITIES.map((k) => [k, mods.output ? has(mods.output, k) : k === "text"])), + interleaved: m.interleaved ?? false, + }, + release_date: m.release_date ?? "", + variants: record(m.variants), + } +} + +// modelOf is a 1.x model, in config or internal form, as a 2.x Model.Info. +export function modelOf(providerID, provider, key, m) { + const caps = record(m.capabilities) + const listFrom = (obj, list) => { + if (obj && typeof obj === "object") return MODALITIES.filter((k) => obj[k]) + if (Array.isArray(list)) return list.map(String) + return ["text"] + } + const input = listFrom(caps.input, m.modalities?.input) + const output = listFrom(caps.output, m.modalities?.output) + const c = record(m.cost) + const cache = record(c.cache) + const npm = m.api?.npm ?? m.provider?.npm + const url = m.api?.url ?? m.provider?.api + const settings = { ...record(json(m.options)) } + if (url && url !== provider.baseURL) settings.baseURL = url + const variants = Object.entries(record(m.variants)) + .filter(([, v]) => !record(v).disabled) + .map(([id, v]) => { + const { disabled: _d, ...rest } = record(json(v)) + return nonEmpty(rest) ? { id, settings: rest } : { id } + }) + const status = ["alpha", "beta", "deprecated", "active"].includes(m.status) ? m.status : "active" + const out = { + id: key, + modelID: String(m.api?.id ?? m.id ?? key), + providerID, + name: String(m.name ?? key), + capabilities: { + tools: caps.toolcall ?? m.tool_call ?? true, + input: input.length ? input : ["text"], + output: output.length ? output : ["text"], + ...(typeof (caps.temperature ?? m.temperature) === "boolean" ? { temperature: caps.temperature ?? m.temperature } : {}), + }, + variants, + time: { released: Date.parse(m.release_date ?? "") || 0 }, + cost: [{ input: money(c.input), output: money(c.output), cache: { read: money(cache.read ?? c.cache_read), write: money(cache.write ?? c.cache_write) } }], + status, + enabled: true, + limit: { context: int(m.limit?.context, 128000) || 128000, output: int(m.limit?.output, 8192) || 8192 }, + } + if (m.limit?.input) out.limit.input = int(m.limit.input) + if (npm && npm !== provider.npm) out.package = packageOf(npm) + if (nonEmpty(settings)) out.settings = settings + const headers = strings(m.headers) + if (nonEmpty(headers)) out.headers = headers + return out +} + +// ---- requests ----------------------------------------------------------- + +const urlOf = (input) => (typeof input === "string" ? input : input instanceof URL ? input.href : input?.url ?? String(input)) + +// wrapFetch sends a request the SDK made from the provider's settings the +// way 1.x would have sent it with the loader's options: the loader's base +// URL in place of the configured one, the account's key in place of the +// placeholder, the loader's headers added, through the loader's fetch. +export function wrapFetch(next, base, options) { + return async (input, init = {}) => { + const o = await options() + let url = urlOf(input) + const from = String(base ?? "").replace(/\/+$/, "") + const to = o.baseURL ? String(o.baseURL).replace(/\/+$/, "") : "" + if (from && to && from !== to && url.startsWith(from)) url = to + url.slice(from.length) + const req = input instanceof Request ? input : null + const headers = new Headers(init.headers ?? req?.headers ?? undefined) + for (const [k, v] of [...headers.entries()]) { + if (!v.includes(PLACEHOLDER)) continue + if (o.key) headers.set(k, v.replaceAll(PLACEHOLDER, o.key)) + else headers.delete(k) + } + for (const [k, v] of Object.entries(strings(o.headers))) headers.set(k, v) + const send = typeof o.fetch === "function" ? o.fetch : next ?? fetch + if (req && init.body === undefined) { + const body = req.method === "GET" || req.method === "HEAD" ? undefined : await req.text() + return send(url, { method: req.method, headers, body, signal: init.signal ?? req.signal }) + } + return send(url, { ...init, headers }) + } +} + +// ---- one plugin --------------------------------------------------------- + +async function setupPlugin(ctx, fn, live) { + const directory = ctx?.location?.directory ?? process.cwd() + // the active account each provider was last read for, and the renewed + // auth a plugin saved for it (2.x has no way for a plugin to write a + // credential back, so it is kept in the plugin's storage, under the + // credential it renews and the token it was renewed from) + const current = new Map() + const saved = new Map() + const signingIn = new Map() + const storeKey = (id, key) => `v1auth/${id}/${key}` + + const client = { + auth: { + async set(arg) { + const id = arg?.path?.id + const body = json(arg?.body) + const cur = id && current.get(id) + // a sign-in's own saves belong to the credential it is about to + // make, not the one that was active when it started + if (!cur || !body || signingIn.get(id) > 0) return { data: true } + const entry = { fp: cur.fp, auth: body } + saved.set(storeKey(id, cur.key), entry) + await Promise.resolve() + .then(() => ctx.storage?.set?.(storeKey(id, cur.key), entry)) + .catch(() => {}) + return { data: true } + }, + async get() { + return { data: undefined } + }, + }, + app: { log: async () => ({ data: true }) }, + } + + const hooks = (await fn({ client, directory, worktree: directory, project: {}, serverUrl: undefined, $: undefined }, {})) ?? {} + const cfg = { provider: {} } + try { + await hooks.config?.(cfg) + } catch {} + const auth = hooks.auth + const authID = auth?.provider ? String(auth.provider) : undefined + const providers = record(cfg.provider) + + // getAuth reads the active account: the credential 2.x holds, or what the + // plugin last saved renewing it. + const getAuth = async (id) => { + const connection = await ctx.integration.connection.active(id).catch(() => undefined) + if (!connection) { + current.delete(id) + return undefined + } + const value = await ctx.integration.connection.resolve(connection).catch(() => undefined) + const base = authOfValue(value) + if (!base) { + current.delete(id) + return undefined + } + const key = connKey(connection) + const fp = fingerprint(base) + // ours: a credential one of this plugin's own sign-ins made + const ours = value?.type === "oauth" && !!record(value.metadata).v1 + current.set(id, { key, fp, ours }) + const sk = storeKey(id, key) + let entry = saved.get(sk) + if (!entry) { + entry = await Promise.resolve() + .then(() => ctx.storage?.get?.(sk)) + .catch(() => undefined) + if (entry) saved.set(sk, entry) + } + return entry && entry.fp === fp && entry.auth ? entry.auth : base + } + + // the sign-in methods + if (authID && Array.isArray(auth.methods) && auth.methods.length) { + const label = (c) => { + const a = authOfValue(c) + const name = a?.name ?? a?.email ?? a?.metadata?.name ?? a?.metadata?.email ?? a?.account + return typeof name === "string" && name ? name : undefined + } + const begin = () => signingIn.set(authID, (signingIn.get(authID) ?? 0) + 1) + const end = () => signingIn.set(authID, Math.max(0, (signingIn.get(authID) ?? 1) - 1)) + const title = record(providers[authID]).name + // 2.x runs a transform again on every reload, from nothing + await ctx.integration.transform((editor) => { + let keyed = false + // the integration 2.x makes for an id it doesn't know is named by the + // id; the provider's own name reads better in the sign-in list + if (typeof title === "string" && title) + try { + editor.update?.(authID, (i) => { + if (i && (!i.name || i.name === authID)) i.name = title + }) + } catch {} + auth.methods.forEach((m, i) => { + if (!m) return + const methodID = `v1-${i}` + const form = fieldsOf(m.prompts) + if (m.type === "api" && typeof m.authorize !== "function") { + if (keyed) return + keyed = true + editor.method.update({ integrationID: authID, method: { type: "key", label: String(m.label ?? "API key"), ...(form ? { form } : {}) } }) + return + } + if (typeof m.authorize !== "function") return + editor.method.update({ + integrationID: authID, + method: { id: methodID, type: "oauth", label: String(m.label ?? authID), ...(form ? { form } : {}) }, + async authorize(answer) { + begin() + let ended = false + const done = () => { + if (!ended) end() + ended = true + } + try { + const inputs = Object.fromEntries(Object.entries(record(answer)).map(([k, v]) => [k, String(v)])) + const r = await m.authorize(inputs) + if (m.type === "api") { + done() + return { url: "", instructions: "", mode: "auto", callback: Promise.resolve().then(() => credentialOf(methodID, r)) } + } + if (r?.method === "code") + return { + url: String(r.url ?? ""), + instructions: String(r.instructions ?? ""), + mode: "code", + callback: async (code) => { + try { + return credentialOf(methodID, await r.callback(code)) + } finally { + done() + } + }, + } + const callback = Promise.resolve() + .then(() => r.callback()) + .then((res) => credentialOf(methodID, res)) + .finally(done) + callback.catch(() => {}) + return { url: String(r?.url ?? ""), instructions: String(r?.instructions ?? ""), mode: "auto", callback } + } catch (e) { + done() + throw e + } + }, + label, + }) + }) + }) + } + + // the loader's options for the account signed in now, read again when + // the account changes + let loaded = null + const loaderOptions = async (id) => { + if (id !== authID || typeof auth?.loader !== "function") { + const a = authID === id ? await getAuth(id) : undefined + return { key: a?.type === "api" ? a.key : undefined } + } + const a = await getAuth(id) + if (!a) return {} + const fp = current.get(id)?.key + "\0" + fingerprint(a) + if (loaded?.fp !== fp) { + const p = Promise.resolve() + .then(() => auth.loader(() => getAuth(id), providerOf(id))) + .then((o) => record(o)) + loaded = { fp, p } + p.catch(() => { + if (loaded?.fp === fp) loaded = null + }) + } + const o = await loaded.p + const key = typeof o.apiKey === "string" && o.apiKey ? o.apiKey : a.type === "api" ? a.key : undefined + return { baseURL: o.baseURL, key, headers: o.headers, fetch: o.fetch } + } + + // the providers, with the list the account's own read last gave each + const source = new Map() + // providerOf is a provider as 1.x hands it to provider.models and + // auth.loader + const providerOf = (id) => { + const conf = record(providers[id]) + const models = Object.fromEntries(Object.entries(record(conf.models)).map(([k, m]) => [k, internalOf(id, conf, k, record(m))])) + return { id, name: conf.name ?? id, source: "config", env: [], options: record(conf.options), models } + } + const listed = new Set() + for (const [id, conf0] of Object.entries(providers)) { + const conf = record(conf0) + // a provider entry that only names the sign-in (no package, no models) + // is not a provider of models + if (!conf.npm && !Object.keys(record(conf.models)).length) continue + const options = record(conf.options) + const { apiKey: _k, baseURL: _b, fetch: _f, headers: optHeaders, ...rest } = options + // with no base URL the config leaves it to the loader: a name that + // can't resolve stands in for it, and each request swaps in the + // loader's (wrapFetch) + const baseURL = options.baseURL ?? conf.api ?? `https://${id.replace(/[^a-z0-9-]/gi, "-")}.opencode2-v1.invalid` + const npm = conf.npm ?? "@ai-sdk/openai-compatible" + const info = { + id, + ...(id === authID ? { integrationID: id } : {}), + name: String(conf.name ?? id), + activation: "auto", + package: packageOf(npm), + settings: { ...record(json(rest)), ...(baseURL ? { baseURL: String(baseURL) } : {}) }, + } + const headers = { ...strings(conf.headers), ...strings(optHeaders) } + if (nonEmpty(headers)) info.headers = headers + const prov = { baseURL, npm } + const fromConfig = () => + Object.entries(record(conf.models)).map(([key, m]) => modelOf(id, prov, key, internalOf(id, conf, key, record(m)))) + listed.add(id) + await ctx.provider.transform((editor) => { + const s = source.get(id) + editor.add({ info, models: s?.models ?? fromConfig(), sourceConnection: s?.connection }) + }) + source.set(id, { conf, prov, models: null, connection: undefined }) + } + + // each request: the placeholder key, and the fetch that swaps in the + // account's. A provider the package doesn't list (one 2.x has built in + // that the plugin only signs in to) is left to 2.x unless the plugin's + // loader serves it and one of the plugin's own sign-ins is active. + const own = authID && !listed.has(authID) && typeof auth?.loader === "function" ? [authID] : [] + for (const id of new Set([...listed, ...own])) { + const base = source.get(id)?.prov.baseURL + await ctx.aisdk.hook( + "sdk", + async (evt) => { + const opts = evt.options + if (!opts || opts.__opencode2) return + if (!listed.has(id)) { + await getAuth(id) + if (!current.get(id)?.ours) return + } + opts.__opencode2 = true + opts.apiKey = PLACEHOLDER + opts.fetch = wrapFetch(opts.fetch, opts.baseURL ?? base, () => loaderOptions(id)) + }, + { providerID: id }, + ) + } + + // refresh reads the account's own list after a sign-in or a switch + const listID = hooks.provider?.id ? String(hooks.provider.id) : undefined + let refreshing = null + let again = false + // the account the list was last read for: a credential.updated that + // leaves it as it was reads nothing again + let readFor = null + const stateOf = async () => { + const c = await ctx.integration.connection.active(authID).catch(() => undefined) + const a = c ? await getAuth(authID) : undefined + return connKey(c) + "\0" + (a ? fingerprint(a) : "") + } + const refresh = () => { + if (refreshing) { + again = true + return refreshing + } + refreshing = (async () => { + try { + if (!authID || !source.has(authID)) return + const s = source.get(authID) + const connection = await ctx.integration.connection.active(authID).catch(() => undefined) + const a = connection ? await getAuth(authID) : undefined + readFor = connKey(connection) + "\0" + (a ? fingerprint(a) : "") + if (!a || listID !== authID || typeof hooks.provider?.models !== "function") { + if (s.connection || s.models) { + s.connection = undefined + s.models = null + await ctx.provider.reload() + } + return + } + const provider = providerOf(authID) + let got + try { + got = await hooks.provider.models(provider, { auth: a }) + } catch { + got = null + } + if (connKey(await ctx.integration.connection.active(authID).catch(() => undefined)) !== connKey(connection)) { + again = true + return + } + s.connection = connection + s.models = got && typeof got === "object" ? Object.entries(got).map(([k, m]) => modelOf(authID, s.prov, k, record(m))) : null + await ctx.provider.reload() + await ctx.model?.reload?.() + } finally { + refreshing = null + if (again && live.on) { + again = false + void refresh() + } + } + })() + return refreshing + } + + if (authID && ctx.event?.subscribe) + void (async () => { + try { + // a switch names the integration; an update (a sign-in, a removal) + // names nothing, so it reads again only when this account changed + const types = ["credential.switched", "credential.updated"] + for await (const event of ctx.event.subscribe({ types })) { + if (!live.on) return + if (event?.type === "credential.switched" && event?.data?.integrationID === authID) { + loaded = null + await refresh() + } else if (event?.type === "credential.updated" && (await stateOf()) !== readFor) { + await refresh() + } + } + } catch {} + })() + // A plugin still reading its list in the background when the first read + // returned (agy-cli: `agy models` and a session probe take seconds, and + // the first read gets its static list) says so through + // provider.warming(): the list is read once more when that ends, once, + // and no later than a minute on. + const warmed = async () => { + let w + try { + w = typeof hooks.provider?.warming === "function" ? hooks.provider.warming() : null + } catch { + return + } + if (!w || typeof w.then !== "function") return + let timer + await Promise.race([ + Promise.resolve(w).catch(() => {}), + new Promise((r) => { + timer = setTimeout(r, 60_000) + timer.unref?.() + }), + ]) + clearTimeout(timer) + if (live.on) await refresh() + } + void refresh().then(warmed, () => {}) +} + +export async function setup(ctx, mod = v1) { + const live = { on: true } + for (const fn of pluginsOf(mod)) { + try { + await setupPlugin(ctx, fn, live) + } catch (e) { + // one plugin failing to start leaves the package's others working + console.error(`${pkg.name}: ${e?.message ?? e}`) + } + } + return () => { + live.on = false + } +} + +export default { id: pkg.name, setup: (ctx) => setup(ctx) } diff --git a/packages/openrouter-decisions/package.json b/packages/openrouter-decisions/package.json new file mode 100644 index 000000000..d52d536de --- /dev/null +++ b/packages/openrouter-decisions/package.json @@ -0,0 +1,19 @@ +{ + "name": "@magpie-community/opencode-openrouter-decisions-auth", + "version": "0.1.0", + "description": "Expose OpenRouter's decision models (Microsoft Decision 1.1/1) through Magpie's System One decision API.", + "type": "module", + "main": "./opencode.mjs", + "exports": {".": "./opencode2.mjs"}, + "files": ["index.mjs", "adapter.mjs", "opencode.mjs", "opencode2.mjs", "README.md", "LICENSE", "NOTICE.md", "CHANGELOG.md", "VERIFICATION.md"], + "scripts": {"test": "node --test test/*.test.mjs"}, + "magpie": { + "name": "OpenRouter Decisions", + "icon": "data:image/svg+xml,%3Csvg%20xmlns%3D%22http%3A%2F%2Fwww.w3.org%2F2000%2Fsvg%22%20viewBox%3D%220%200%2064%2064%22%3E%3Ccircle%20cx%3D%2232%22%20cy%3D%2232%22%20r%3D%2229%22%20fill%3D%22%230f172a%22%2F%3E%3Cpath%20d%3D%22M20%2040a12%2012%200%201%201%2012%2012%22%20fill%3D%22none%22%20stroke%3D%22%2365a3ff%22%20stroke-width%3D%225%22%20stroke-linecap%3D%22round%22%2F%3E%3Ccircle%20cx%3D%2244%22%20cy%3D%2228%22%20r%3D%226%22%20fill%3D%22%2365a3ff%22%2F%3E%3C%2Fsvg%3E", + "maxConcurrency": 8 + }, + "engines": {"node": ">=20"}, + "license": "MIT", + "keywords": ["opencode", "opencode-plugin", "magpie", "magpie-plugin", "openrouter", "decisions", "system-one"], + "repository": {"type": "git", "url": "https://github.com/magpie-community/plugins.git", "directory": "packages/openrouter-decisions"} +} diff --git a/packages/openrouter-decisions/test/adapter.test.mjs b/packages/openrouter-decisions/test/adapter.test.mjs new file mode 100644 index 000000000..4b45ef80d --- /dev/null +++ b/packages/openrouter-decisions/test/adapter.test.mjs @@ -0,0 +1,186 @@ +import assert from "node:assert/strict"; +import test from "node:test"; +import { createDecisionFetch, fromOpenRouterResponse, MODELS, toOpenRouterRequest } from "../adapter.mjs"; + +const UPSTREAM_MODEL = MODELS["microsoft-decision-1.1"].upstream; + +const body = () => ({ + model: "microsoft-decision-1.1", + state: { message: "What does HTTP 404 mean?", previous_message_kind: "question" }, + questions: { + intent: { type: "choice", instructions: "Which task?", criteria: { question: "Explain a fact", edit: null } }, + effort: { type: "score", instructions: "How hard?", criteria: ["Little work", "Some work", "Much work", "Most work"] }, + levels: { type: "noul", instructions: "Are these kinds levels of one scale?" }, + }, +}); + +const reply = () => ({ + id: "gen-dec-test", + model: UPSTREAM_MODEL, + provider: "Microsoft", + usage: { input_tokens: 42, output_tokens: 0, cost: 0.0000018 }, + answers: { + levels: { type: "noul", noul: 0.9, confidence: 0.9 }, + intent: { type: "choice", choice: "question", confidence: 0.8, probabilities: { question: 0.9, edit: 0.1 } }, + effort: { type: "score", score: 1.84, confidence: 0.85, probabilities: { 0: 0.01, 1: 0.14, 2: 0.85, 3: 0 } }, + }, +}); + +test("serves both Microsoft decision models and maps the upstream slug", () => { + for (const id of Object.keys(MODELS)) { + const out = toOpenRouterRequest({ ...body(), model: id }); + assert.equal(out.model, MODELS[id].upstream); + } + assert.throws(() => toOpenRouterRequest({ ...body(), model: "microsoft-decision-9" })); +}); + +test("maps the System One request onto the OpenRouter Decisions request", () => { + const out = toOpenRouterRequest(body()); + assert.equal(out.model, UPSTREAM_MODEL); + assert.equal(typeof out.state, "string"); + assert.deepEqual(JSON.parse(out.state), body().state); + assert.deepEqual(out.questions.intent, { type: "choice", instructions: "Which task?", criteria: { question: "Explain a fact", edit: "edit" } }); + assert.deepEqual(out.questions.effort.criteria, ["Little work", "Some work", "Much work", "Most work"]); + assert.deepEqual(out.questions.levels, { type: "noul", instructions: "Are these kinds levels of one scale?" }); +}); + +test("passes text state through and expands the choices shorthand", () => { + const out = toOpenRouterRequest({ model: "microsoft-decision-1.1", state: "hello", questions: { intent: { type: "choice", instructions: "Kind?", choices: ["question", "edit"] } } }); + assert.equal(out.state, "hello"); + assert.deepEqual(out.questions.intent.criteria, { question: "question", edit: "edit" }); +}); + +test("keeps multimodal state parts as an array of strings", () => { + const state = ["Look:", { type: "image_url", image_url: { url: "data:image/jpeg;base64,AA" } }]; + const out = toOpenRouterRequest({ ...body(), state }); + assert.equal(Array.isArray(out.state), true); + assert.equal(out.state[0], "Look:"); + assert.deepEqual(JSON.parse(out.state[1]), state[1]); +}); + +test("supplies fallback instructions and keeps predicate definitions", () => { + const source = body(); + for (const question of Object.values(source.questions)) question.instructions = " "; + source.questions.levels.criteria = { true: "Matches", false: "Does not match" }; + const out = toOpenRouterRequest(source); + for (const question of Object.values(out.questions)) assert.ok(question.instructions.trim().length > 0); + assert.deepEqual(out.questions.levels.criteria, { true: "Matches", false: "Does not match" }); +}); + +test("restores probabilities, confidence, legend and usage", () => { + const out = fromOpenRouterResponse(reply(), body()); + assert.equal(out.answers.intent.choice, "question"); + assert.deepEqual(out.answers.intent.probabilities, { question: 0.9, edit: 0.1 }); + assert.equal(out.answers.effort.score, 1.84); + assert.deepEqual(out.answers.effort.probabilities, { 0: 0.01, 1: 0.14, 2: 0.85, 3: 0 }); + assert.deepEqual(out.answers.levels, { type: "noul", noul: 0.9 }); + assert.deepEqual(out.usage, reply().usage); +}); + +test("a per-question refusal stays a refusal", () => { + const upstream = reply(); + upstream.answers.intent = { type: "refusal" }; + const out = fromOpenRouterResponse(upstream, body()); + assert.deepEqual(out.answers.intent, { type: "refusal" }); + assert.equal(out.answers.effort.score, 1.84); +}); + +test("rejects unsupported requests before calling upstream", () => { + for (const source of [ + { ...body(), stream: true }, + { ...body(), model: "typesafe/jev-1.14" }, + { ...body(), questions: {} }, + { ...body(), questions: { x: { type: "text", instructions: "Explain" } } }, + { ...body(), questions: { x: { type: "choice", instructions: "Kind?", criteria: { only: null } } } }, + { ...body(), questions: { x: { ...body().questions.intent, choices: ["a", "b"] } } }, + { ...body(), questions: { x: { type: "noul", criteria: { other: "x" } } } }, + ]) assert.throws(() => toOpenRouterRequest(source)); +}); + +test("rejects missing, wrong-type and malformed answers", () => { + for (const mutate of [ + (r) => delete r.answers.intent, + (r) => { r.answers.intent.type = "noul"; }, + (r) => { r.answers.intent.choice = "unknown"; }, + (r) => { r.answers.intent.probabilities = { question: 1 }; }, + (r) => { r.answers.intent.probabilities = { question: 0.5, edit: 0.4 }; }, + (r) => { r.answers.intent.confidence = 2; }, + (r) => { r.answers.effort.score = 9; }, + (r) => { r.answers.levels.noul = -0.1; }, + ]) { + const upstream = reply(); + mutate(upstream); + assert.throws(() => fromOpenRouterResponse(upstream, body())); + } +}); + +test("fetch posts to the fixed Decisions URL with the current key only", async () => { + let call; + const fetcher = createDecisionFetch({ getKey: () => "sk-or-fixture", fetchImpl: async (url, init) => { call = { url, init }; return Response.json(reply(), { headers: { "x-request-id": "req-test", "set-cookie": "no" } }); } }); + const response = await fetcher("https://openrouter.ai/api/v1/systemone", { method: "POST", headers: { authorization: "Bearer gateway-key", "x-magpie-private": "no" }, body: JSON.stringify(body()) }); + assert.equal(response.status, 200); + assert.equal(call.url, "https://openrouter.ai/api/alpha/decisions"); + assert.equal(call.init.redirect, "error"); + assert.equal(call.init.headers.get("authorization"), "Bearer sk-or-fixture"); + assert.equal(call.init.headers.has("x-magpie-private"), false); + assert.equal(response.headers.has("set-cookie"), false); + assert.equal(response.headers.get("x-request-id"), "req-test"); + assert.equal((await response.json()).answers.intent.choice, "question"); +}); + +test("does not send malformed requests or arbitrary endpoints upstream", async () => { + let sent = 0; + const fetcher = createDecisionFetch({ getKey: () => "fixture", fetchImpl: async () => { sent++; throw Error("should not send"); } }); + for (const [url, init] of [ + ["https://elsewhere.example/v1/systemone", { method: "POST", body: JSON.stringify(body()) }], + ["https://openrouter.ai/api/v1/chat/completions", { method: "POST", body: JSON.stringify(body()) }], + ["https://openrouter.ai/api/v1/systemone", { method: "GET" }], + ["https://openrouter.ai/api/v1/systemone", { method: "POST", body: "nope" }], + ]) assert.equal((await fetcher(url, init)).status, 400); + assert.equal(sent, 0); +}); + +test("missing credentials become 503 and never reach upstream", async () => { + let sent = 0; + const fetcher = createDecisionFetch({ getKey: async () => undefined, fetchImpl: async () => { sent++; return Response.json({}); } }); + assert.equal((await fetcher("https://openrouter.ai/api/v1/systemone", { method: "POST", body: JSON.stringify(body()) })).status, 503); + assert.equal(sent, 0); +}); + +test("preserves upstream error status, body and retry headers", async () => { + for (const status of [401, 429, 503]) { + const raw = JSON.stringify({ error: { message: "upstream refused" } }); + const fetcher = createDecisionFetch({ getKey: () => "fixture", fetchImpl: async () => new Response(raw, { status, headers: { "retry-after": "5", "content-type": "application/json" } }) }); + const response = await fetcher("https://openrouter.ai/api/v1/systemone", { method: "POST", body: JSON.stringify(body()) }); + assert.equal(response.status, status); + assert.equal(await response.text(), raw); + assert.equal(response.headers.get("retry-after"), "5"); + } +}); + +test("a transient network failure is retried once, an HTTP error is not", async () => { + let calls = 0; + const retryFetcher = createDecisionFetch({ getKey: () => "fixture", fetchImpl: async () => { calls++; if (calls === 1) throw new Error("socket hang up"); return Response.json(reply()); } }); + const ok = await retryFetcher("https://openrouter.ai/api/v1/systemone", { method: "POST", body: JSON.stringify(body()) }); + assert.equal(ok.status, 200); + assert.equal(calls, 2); + const noRetryFetcher = createDecisionFetch({ getKey: () => "fixture", fetchImpl: async () => { calls++; return new Response("{}", { status: 503 }); } }); + const refused = await noRetryFetcher("https://openrouter.ai/api/v1/systemone", { method: "POST", body: JSON.stringify(body()) }); + assert.equal(refused.status, 503); + assert.equal(calls, 3); + const giveUpFetcher = createDecisionFetch({ getKey: () => "fixture", fetchImpl: async () => { throw new Error("ECONNREFUSED"); } }); + const failed = await giveUpFetcher("https://openrouter.ai/api/v1/systemone", { method: "POST", body: JSON.stringify(body()) }); + assert.equal(failed.status, 502); + assert.match(JSON.stringify(await failed.json()), /ECONNREFUSED/); +}); + +test("an incomplete upstream success becomes 502", async () => { + const fetcher = createDecisionFetch({ getKey: () => "fixture", fetchImpl: async () => Response.json({ model: UPSTREAM_MODEL, answers: {} }) }); + assert.equal((await fetcher("https://openrouter.ai/api/v1/systemone", { method: "POST", body: JSON.stringify(body()) })).status, 502); +}); + +test("cancellation reaches upstream", async () => { + const controller = new AbortController(); + const fetcher = createDecisionFetch({ getKey: () => "fixture", fetchImpl: async (_url, init) => { assert.equal(init.signal.aborted, false); controller.abort(); init.signal.throwIfAborted(); } }); + await assert.rejects(fetcher("https://openrouter.ai/api/v1/systemone", { method: "POST", signal: controller.signal, body: JSON.stringify(body()) }), { name: "AbortError" }); +}); diff --git a/registry.json b/registry.json index 8237e459e..c72255ec8 100644 --- a/registry.json +++ b/registry.json @@ -1,204 +1,1131 @@ { "version": 1, "plugins": [ - { "package": "@magpie-community/opencode-gumloop-auth", "name": "Gumloop", "icon": "https://www.gumloop.com/favicon.ico", "providers": ["gumloop"], "community": true, - "summary": { "en": "Gumloop with your API key and user ID: account models, streaming, tool calls and a locally calibrated credit estimate.", "zh": "使用 API key 和用户 ID 接入 Gumloop:按账号发现模型,支持流式响应、工具调用和本地校准的积分估算。" } }, - { "package": "@magpie-community/opencode-agnes-auth", "name": "Agnes AI", "icon": "https://www.agnes-ai.com/images/biglogo.png", "providers": ["agnes"], "community": true, - "summary": { "en": "Agnes AI international with Chrome browser sign-in: live OpenAI-compatible text models, Token Plan and Agent Plan usage, and the USD wallet balance.", "zh": "通过 Chrome 浏览器登录 Agnes AI 国际版:动态发现 OpenAI 兼容文本模型,显示 Token Plan、Agent Plan 用量与美元余额,无需手动复制会话 token。" } }, - { "package": "@magpie-community/opencode-atomcode-auth", "name": "AtomCode", "icon": "https://atomgit.com/favicon.ico", "providers": ["atomcode"], "community": true, - "summary": { "en": "Unofficial AtomGit CodingPlan access for magpie and OpenCode: browser OAuth or local AtomCode CLI sign-in, signed OpenAI-compatible requests, live models, quota windows, and a read-only diagnostics page.", "zh": "非官方接入 AtomGit CodingPlan:支持浏览器 OAuth 或复用本机 AtomCode CLI 登录,发送签名请求,显示实时模型与额度窗口,并提供只读诊断页。" } }, - { "package": "@magpie-community/gateway-powercontext", "name": "PowerContext", "kind": "middleware", "community": true, + { + "package": "@magpie-community/opencode-gumloop-auth", + "name": "Gumloop", + "icon": "https://www.gumloop.com/favicon.ico", + "providers": [ + "gumloop" + ], + "community": true, + "summary": { + "en": "Gumloop with your API key and user ID: account models, streaming, tool calls and a locally calibrated credit estimate.", + "zh": "使用 API key 和用户 ID 接入 Gumloop:按账号发现模型,支持流式响应、工具调用和本地校准的积分估算。" + } + }, + { + "package": "@magpie-community/opencode-agnes-auth", + "name": "Agnes AI", + "icon": "https://www.agnes-ai.com/images/biglogo.png", + "providers": [ + "agnes" + ], + "community": true, + "summary": { + "en": "Agnes AI international with Chrome browser sign-in: live OpenAI-compatible text models, Token Plan and Agent Plan usage, and the USD wallet balance.", + "zh": "通过 Chrome 浏览器登录 Agnes AI 国际版:动态发现 OpenAI 兼容文本模型,显示 Token Plan、Agent Plan 用量与美元余额,无需手动复制会话 token。" + } + }, + { + "package": "@magpie-community/opencode-atomcode-auth", + "name": "AtomCode", + "icon": "https://atomgit.com/favicon.ico", + "providers": [ + "atomcode" + ], + "community": true, + "summary": { + "en": "Unofficial AtomGit CodingPlan access for magpie and OpenCode: browser OAuth or local AtomCode CLI sign-in, signed OpenAI-compatible requests, live models, quota windows, and a read-only diagnostics page.", + "zh": "非官方接入 AtomGit CodingPlan:支持浏览器 OAuth 或复用本机 AtomCode CLI 登录,发送签名请求,显示实时模型与额度窗口,并提供只读诊断页。" + } + }, + { + "package": "@magpie-community/gateway-powercontext", + "name": "PowerContext", + "kind": "middleware", + "community": true, "icon": "https://raw.githubusercontent.com/oceanbase/powercontext/master/src/powercontext/server/dashboard/static/brand/favicon-color.svg", - "summary": { "en": "Recall from your PowerContext service and optionally save complete conversations. Configure the service URL first; capture is off by default.", "zh": "从自己的 PowerContext 服务召回记忆,并按需存回完整对话。先配置服务地址;对话存回默认关闭。" } }, - { "package": "@magpie-community/gateway-nowledge-mem", "name": "Nowledge Mem", "kind": "middleware", "community": true, - "summary": { "en": "Give every harness behind Magpie Nowledge Mem automatically: recall and the Context Bundle (or Working Memory) before each request, per-harness provenance, optional import of completed conversations as Mem Threads, and a real connection test in the page. A local or remote Mem; saving conversations is off by default.", "zh": "让 Magpie 后面的每个 harness 自动接入 Nowledge Mem:每次请求前召回记忆与 Context Bundle(缺失时用 Working Memory),记录来自哪个 harness,按需把已完成的对话导入为 Mem Thread,并在页面中提供真实的连接测试。本机或远端 Mem 都可用;对话导入默认关闭。" } }, - { "package": "@magpie-community/gateway-aelios", "name": "Aelios", "kind": "middleware", "community": true, - "summary": { "en": "Long-term memory from your own Aelios on Cloudflare: recall before each turn for every agent behind magpie, and optionally record the conversation so Aelios can learn it. Set it up on the plugin's own page, with a connection test and per-model or per-agent routing to separate assistants. Recording is off by default.", "zh": "接入你自己部署在 Cloudflare 上的 Aelios 长期记忆:magpie 后面的每个 agent 每轮自动召回,按需把对话交给 Aelios 记住。在插件自己的页面里配置,可选择连接测试,并按模型或客户端分流到不同的助手。记录默认关闭。" } }, - { "package": "@magpie-community/gateway-respire-memory", "name": "respire-memory", "kind": "middleware", "community": true, + "summary": { + "en": "Recall from your PowerContext service and optionally save complete conversations. Configure the service URL first; capture is off by default.", + "zh": "从自己的 PowerContext 服务召回记忆,并按需存回完整对话。先配置服务地址;对话存回默认关闭。" + } + }, + { + "package": "@magpie-community/gateway-nowledge-mem", + "name": "Nowledge Mem", + "kind": "middleware", + "community": true, + "summary": { + "en": "Give every harness behind Magpie Nowledge Mem automatically: recall and the Context Bundle (or Working Memory) before each request, per-harness provenance, optional import of completed conversations as Mem Threads, and a real connection test in the page. A local or remote Mem; saving conversations is off by default.", + "zh": "让 Magpie 后面的每个 harness 自动接入 Nowledge Mem:每次请求前召回记忆与 Context Bundle(缺失时用 Working Memory),记录来自哪个 harness,按需把已完成的对话导入为 Mem Thread,并在页面中提供真实的连接测试。本机或远端 Mem 都可用;对话导入默认关闭。" + } + }, + { + "package": "@magpie-community/gateway-aelios", + "name": "Aelios", + "kind": "middleware", + "community": true, + "summary": { + "en": "Long-term memory from your own Aelios on Cloudflare: recall before each turn for every agent behind magpie, and optionally record the conversation so Aelios can learn it. Set it up on the plugin's own page, with a connection test and per-model or per-agent routing to separate assistants. Recording is off by default.", + "zh": "接入你自己部署在 Cloudflare 上的 Aelios 长期记忆:magpie 后面的每个 agent 每轮自动召回,按需把对话交给 Aelios 记住。在插件自己的页面里配置,可选择连接测试,并按模型或客户端分流到不同的助手。记录默认关闭。" + } + }, + { + "package": "@magpie-community/gateway-respire-memory", + "name": "respire-memory", + "kind": "middleware", + "community": true, "icon": "https://raw.githubusercontent.com/risense-ai/respire-magpie/v0.3.0/page/icon.svg", - "summary": { "en": "Help AI agents evolve through experience: per-turn recall and complete engine-policy reads, automatic process diaries and grounded archival with write readback. Configure an existing local Runtime and generation model first. Archival defaults on; raw capture off. Background storage happens after reply completion.", "zh": "让 AI Agent 通过经验积累持续自进化:每轮召回并读完整政策,自动保存过程日记与整理长期记忆,写后回读。先配置已有本地 Runtime 和生成模型;归档默认开启,原始采集关闭。宿主限制下存储在回复完成后执行。" } }, - { "package": "@openviking/magpie-plugin", "name": "OpenViking", "kind": "middleware", "community": true, + "summary": { + "en": "Help AI agents evolve through experience: per-turn recall and complete engine-policy reads, automatic process diaries and grounded archival with write readback. Configure an existing local Runtime and generation model first. Archival defaults on; raw capture off. Background storage happens after reply completion.", + "zh": "让 AI Agent 通过经验积累持续自进化:每轮召回并读完整政策,自动保存过程日记与整理长期记忆,写后回读。先配置已有本地 Runtime 和生成模型;归档默认开启,原始采集关闭。宿主限制下存储在回复完成后执行。" + } + }, + { + "package": "@openviking/magpie-plugin", + "name": "OpenViking", + "kind": "middleware", + "community": true, "icon": "https://raw.githubusercontent.com/volcengine/OpenViking/main/docs/images/ov-logo-icon.png", - "summary": { "en": "Recall OpenViking context before model requests and optionally capture completed conversations. Reuses your existing OpenViking configuration; capture is off by default. Maintained by the OpenViking project.", "zh": "在模型请求前召回 OpenViking 上下文,并按需采集完整对话。复用现有 OpenViking 配置;采集默认关闭。由 OpenViking 项目维护。" } }, - { "package": "@vshulcz/magpie-deja", "name": "deja-vu", "kind": "middleware", "community": true, + "summary": { + "en": "Recall OpenViking context before model requests and optionally capture completed conversations. Reuses your existing OpenViking configuration; capture is off by default. Maintained by the OpenViking project.", + "zh": "在模型请求前召回 OpenViking 上下文,并按需采集完整对话。复用现有 OpenViking 配置;采集默认关闭。由 OpenViking 项目维护。" + } + }, + { + "package": "@vshulcz/magpie-deja", + "name": "deja-vu", + "kind": "middleware", + "community": true, "icon": "https://raw.githubusercontent.com/vshulcz/deja-vu/main/docs/assets/icon.svg", - "summary": { "en": "Recall local session history from the transcripts Claude Code, Codex, Hermes, Gemini CLI, OpenCode, Cursor and DSH write on this machine before each request. Scopes lookup to the request's project directory (ctx.cwd) by default and scrubs secret patterns before attaching recalled excerpts to the request sent to your model provider. When recalled history helped, the model is asked to end its reply with a one-line \"déjà vu: … — reusing it.\" note.", "zh": "每次请求前从本机 Claude Code、Codex、Hermes、Gemini CLI、OpenCode、Cursor 与 DSH 的磁盘会话记录中召回相关历史。默认严格限定在请求所属的项目目录(ctx.cwd)内检索,并对密钥与令牌执行脱敏后,将召回片段随请求一并发送给所配置的模型供应商。当召回的历史记录有帮助时,会请模型在回复末尾附上一行“déjà vu: … — reusing it.”说明。" } }, - { "package": "@magpie-community/opencode-copilot-auth", "name": "GitHub Copilot", "icon": "githubcopilot", "providers": ["github-copilot"], "community": true, - "summary": { "en": "GitHub Copilot on GitHub.com or GitHub Enterprise, with premium request or credit usage, remaining allowance and quota reset dates.", "zh": "GitHub Copilot:支持 GitHub.com 与 GitHub Enterprise,显示高级请求或积分用量、剩余额度和重置时间。" } }, - { "package": "@magpie-community/opencode-joycode-auth", "name": "JoyCode", "providers": ["joycode"], "community": true, - "summary": { "en": "Unofficial JoyCode provider: JD account browser sign-in, live models and credit multipliers, plan credits and streaming tool calls.", "zh": "非官方 JoyCode 插件:京东账号浏览器登录,动态模型与积分倍率、套餐额度及流式工具调用。" } }, - {"package": "@magpie-community/opencode-doubao-work-auth", "name": "豆包工作", "icon": "doubao-color", "providers": ["doubao-work"], "community": true, "summary": {"en": "Not a model API: Doubao Work has none for other clients. This drives its desktop app (or Doubao's) the way a person does, so each request is one run of Doubao's cloud agent in a new task, with the client's tools asked for as text. macOS and Windows; unofficial.", "zh": "不是模型接口:豆包工作没有给第三方客户端用的模型接口。本插件像人一样操作豆包工作(或豆包)桌面版,每个请求是豆包云端智能体在一个新任务里的一次执行,客户端的工具以文本方式调用。macOS 和 Windows;非官方。"}}, - {"package": "@magpie-community/opencode-kimi-work-auth", "name": "Kimi Work", "icon": "kimi", "providers": ["kimi-work"], "community": true, "summary": {"en": "Kimi desktop Work subscription through its local sign-in, with live Work models and shared membership usage. Separate from Kimi Code.", "zh": "使用本机普通 Kimi 桌面版的 Work 登录,自动读取 Work 模型和会员共享用量;独立于 Kimi Code。"}}, - { "package": "@magpie-community/opencode-zen-free-auth", "name": "OpenCode Zen Free", "icon": "opencode", "providers": ["opencode-zen-free"], "community": true, - "summary": { "en": "OpenCode Zen's free models with the public credential, automatic discovery and native streaming protocols.", "zh": "使用 public 凭证接入 OpenCode Zen 免费模型,自动发现模型并支持原生流式协议。" } }, - { "package": "@magpie-community/opencode-zcode-auth", "name": "ZCode", "icon": "zcode", "providers": ["zcode"], "community": true, "replaces": "zcode", - "summary": { "en": "Z.ai and BigModel GLM Coding Plan, team seats and ZCode's free Start Plan, with a daily claim of its gift plans.", "zh": "Z.ai 与智谱 GLM Coding Plan,支持团队席位和 ZCode 免费 Start Plan,可每日领取 ZCode 活动包。" } }, - { "package": "@magpie-community/opencode-qoder-auth", "name": "Qoder", "icon": "qoder", "providers": ["qoder", "qoder-cn", "qwenwork", "qwenwork-enterprise"], "community": true, "replaces": "qoder", - "summary": { "en": "Your Qoder subscription: international (qoder.com) or Qoder CN, the China site (qoder.cn, Alibaba Cloud or phone sign-in), each its own provider. Also QwenWork (千问办公): the public cloud or the enterprise flagship with its own domain. Unofficial for QwenWork.", "zh": "用 Qoder 自己的设备登录使用 Qoder 订阅:国际版(qoder.com)与 Qoder 中国版(Qoder CN,qoder.cn,阿里云账号 / 手机号登录)各是一个订阅;也接入千问办公:公有云与企业旗舰版(输入企业自己的域名)。千问办公部分非官方。" } }, - { "package": "@magpie-community/opencode-devin-auth", "name": "Devin", "icon": "devin", "providers": ["devin"], "community": true, "replaces": "devin", - "summary": { "en": "Devin's models with your Devin account — a browser sign-in or the devin CLI's.", "zh": "用 Devin 账号使用 Devin 的模型:浏览器登录,或沿用 devin CLI 的登录。" } }, - { "package": "@magpie-community/opencode-zed-auth", "name": "Zed", "icon": "zed", "providers": ["zed"], "community": true, "replaces": "zed", - "summary": { "en": "Claude, GPT, Gemini and Grok hosted by Zed, on your Zed plan.", "zh": "用 Zed 订阅使用 Zed 托管的 Claude、GPT、Gemini 和 Grok。" } }, - { "package": "@magpie-community/opencode-factory-auth", "name": "Factory", "icon": "factory", "providers": ["factory"], "community": true, "replaces": "factory", - "summary": { "en": "Factory's Droid subscription: models from Anthropic, OpenAI, xAI and open labs.", "zh": "Factory Droid 订阅:Anthropic、OpenAI、xAI 及开源模型。" } }, - { "package": "@magpie-community/opencode-grok-auth", "name": "Grok", "icon": "xai", "providers": ["grok"], "community": true, "replaces": "grok", - "summary": { "en": "SuperGrok and X Premium+, through the Grok Build CLI's sign-in.", "zh": "通过 Grok Build CLI 的登录使用 SuperGrok 和 X Premium+。" } }, - { "package": "@magpie-community/opencode-commandcode-auth", "name": "Command Code", "icon": "commandcode", "providers": ["commandcode-plan"], "community": true, "replaces": "commandcode-plan", - "summary": { "en": "Command Code plans — Pro, GOAT, Max, Ultra, Go and Teams.", "zh": "Command Code 各档套餐:Pro、GOAT、Max、Ultra、Go 和 Teams。" } }, - { "package": "@magpie-community/opencode-mimo-auth", "name": "Xiaomi MiMo", "icon": "mimocode", "providers": ["mimo-app"], "community": true, "replaces": "mimo-app", - "summary": { "en": "Xiaomi MiMo models with your Xiaomi account.", "zh": "用小米账号使用小米 MiMo 模型。" } }, - { "package": "@magpie-community/opencode-cursor-auth", "name": "Cursor", "icon": "cursor", "providers": ["cursor"], "community": true, "replaces": "cursor", - "summary": { "en": "Cursor Pro, Pro+, Ultra and Teams — every model your plan has, with tool calls.", "zh": "Cursor Pro、Pro+、Ultra 与 Teams:套餐里的所有模型,支持工具调用。" } }, - { "package": "@magpie-community/opencode-kiro-auth", "name": "Kiro", "icon": "kiro-color", "providers": ["kiro"], "community": true, "replaces": "kiro", - "summary": { "en": "Kiro Free, Pro, Pro+ and Power — Kiro's sign-in, kiro-cli's, an API key, or accounts imported by refresh token.", "zh": "Kiro Free、Pro、Pro+ 与 Power:Kiro 登录、沿用 kiro-cli 的登录、API key,或按 refresh token 导入账号。" } }, - { "package": "@magpie-community/opencode-minimax-auth", "name": "MiniMax Code", "icon": "minimax-color", "providers": ["minimax-code", "minimax-code-global"], "community": true, - "summary": { "en": "MiniMax Code credits and M Plan with device sign-in, native web search, image/video generation and media tools through Magpie MCP.", "zh": "MiniMax Code 积分与 M Plan、设备登录,提供原生网络搜索、生图、生视频及多模态 MCP 工具。" } }, - { "package": "@magpie-community/opencode-warp-auth", "name": "Warp", "icon": "warp", "providers": ["warp"], "community": true, - "summary": { "en": "Warp (the terminal) AI request credits through the Warp app's sign-in: the auto routers and every model family the account lists.", "zh": "用 Warp 终端应用的登录使用 Warp AI 请求额度:auto 路由与账户可用的全部模型家族。" } }, - { "package": "@magpie-community/opencode-workbuddy-auth", "name": "WorkBuddy", "icon": "workbuddy-color", "providers": ["workbuddy", "workbuddy-ai"], "community": true, "replaces": "workbuddy", - "summary": { "en": "WorkBuddy's China build (CodeBuddy plan) and its international build.", "zh": "WorkBuddy 国内版(CodeBuddy 套餐)与国际版。" } }, - { "package": "@magpie-community/opencode-cline-auth", "name": "Cline", "icon": "cline", "providers": ["cline"], "community": true, - "summary": { "en": "Cline's models on your own Cline account — usage billing, ClinePass and Cline's free models — with Cline's device sign-in or an API key.", "zh": "用自己的 Cline 账号使用 Cline 的模型——用量计费、ClinePass 与免费模型——支持 Cline 设备登录或 API key。" } }, - { "package": "@magpie-community/opencode-trae-auth", "name": "Trae", "icon": "trae", "providers": ["trae-cn", "trae-global"], "community": true, - "summary": { "en": "Trae CN (trae.cn) and Trae international (trae.ai) models with your own account, free tier included — Trae's own browser sign-in. Experimental.", "zh": "用 Trae CN(trae.cn)与 Trae 国际版(trae.ai)账号使用各自的模型,含免费版,使用 Trae 自己的浏览器登录。实验性。" } }, - { "package": "@magpie-community/opencode-opendesign-auth", "name": "Open Design", "providers": ["opendesign"], "community": true, - "summary": { "en": "Open Design Cloud's Design Plan (open-design.ai): the account's device sign-in, the models the plan entitles it to, and the plan's 5-hour and 7-day allowance and wallet balance.", "zh": "Open Design Cloud 的 Design Plan(open-design.ai):账号设备登录、套餐可用的模型,以及 5 小时与 7 天额度、钱包余额。" } }, - { "package": "@magpie-community/opencode-catpaw-auth", "name": "CatPaw", "providers": ["catpaw"], "community": true, - "summary": { "en": "Meituan's 妙手 CatPaw (ai.catpaw.meituan.com): a browser web sign-in or the desktop client's saved credentials, chat through its round/event/turn conversation protocol, models and credit balance.", - "zh": "美团妙手 CatPaw(ai.catpaw.meituan.com):浏览器网页登录或妙手桌面端已保存的登录态,经其 round/event/turn 会话协议对话,含模型与积分余额。" } }, - { "package": "@magpie-community/opencode-loomy-auth", "name": "Loomy", "providers": ["loomy"], "community": true, - "summary": { "en": "iFlytek's Loomy (loomyad.xunfei.cn): SMS sign-in, OpenAI-compatible chat with reasoning-effort folding and image markers, the live model catalog, two point wallets and the daily first-login grant.", - "zh": "讯飞 Loomy(loomyad.xunfei.cn):短信验证码登录,OpenAI 兼容对话(思考档折叠与图片标记适配),实时模型目录、双积分钱包与每日首登赠送。" } }, - { "package": "@magpie-community/opencode-nous-auth", "name": "Nous Portal (OAuth)", "icon": "nousresearch", "providers": ["nous-portal"], "community": true, - "summary": { "en": "Nous Portal models, with Nous Portal's device sign-in.", "zh": "通过 Nous Portal 自己的设备码登录使用 Nous Portal 模型。" } }, - { "package": "@magpie-community/opencode-raccoon-auth", "name": "商汤小浣熊", "icon": "sensenova-color", "providers": ["raccoon"], "community": true, - "summary": { "en": "SenseTime Raccoon (xiaohuanxiong.com) models, account credits and daily check-in, with browser sign-in or Raccoon desktop's live sign-in.", "zh": "商汤小浣熊(xiaohuanxiong.com)模型、积分余额与每日签到,支持浏览器登录或直接复用小浣熊桌面端登录态。" } }, - { "package": "@magpie-community/opencode-sensenova-auth", "name": "SenseNova", "icon": "sensenova-color", "providers": ["sensenova-token-plan"], "community": true, - "summary": { "en": "SenseNova (SenseTime 日日新) Token Plan: a browser sign-in (or the console account and password) reads the credit pools, their 5-hour and 7-day windows and the expiring grants, and serves the plan's own models — the browser sign-in reuses or creates a Token Plan API key for you.", "zh": "商汤日日新 Token Plan:浏览器登录(或控制台账号密码)读取积分池、5 小时 / 7 天窗口与将过期的返赠余额,并直接调用计划内模型——浏览器登录会自动复用或创建 Token Plan API Key。" } }, - { "package": "@magpie-community/opencode-inkstone-auth", "name": "书生·端砚", "icon": "https://raw.githubusercontent.com/magpie-community/plugins/main/packages/inkstone/icon.svg", "providers": ["intern-inkstone"], "community": true, - "summary": { "en": "Unofficial: Intern InkStone(书生·端砚) Token Plan: a browser sign-in reads the ink-credit balance, the 5-hour and 7-day windows and the credit packs, and creates an ink-* API key (or reuses a pasted one) to serve the plan's models for chat.", "zh": "非官方:书生·端砚(Intern InkStone)墨点计划:浏览器登录读取墨点余额、5 小时 / 7 天窗口与墨点包,并创建 ink-* API Key(或复用现有 Key)直接调用计划内模型对话。" } }, - { "package": "@magpie-community/opencode-autoclaw-auth", "name": "AutoClaw", "providers": ["autoclaw", "autoclaw-cn"], "community": true, - "summary": { "en": "AutoClaw (Zhipu OpenClaw) cloud accounts: oversea web sign-in (Google/z.ai) or CN phone SMS, credit balance and daily check-in.", "zh": "AutoClaw(智谱澳龙/OpenClaw)云端账号:国际版网页登录(Google/z.ai)或国内手机验证码,积分余额与每日签到。" } }, - { "package": "@magpie-community/opencode-codearts-auth", "name": "CodeArts / OfficeAce", "providers": ["codearts", "officeace"], "community": true, - "summary": { "en": "Unofficial: Huawei Cloud CodeArts (码道) coding plans and OfficeAce (AgentArts), each signed in with Huawei Cloud OAuth on its own card — GLM, DeepSeek, openPangu and OfficeAce's MaaS models, live models and credit rates, credit usage by bucket and the daily check-in. Replaces officeace-auth; accounts signed in there keep working.", "zh": "非官方:华为云码道(CodeArts)编码套餐与 OfficeAce(AgentArts),两张卡各自用华为云 OAuth 登录;GLM、DeepSeek、openPangu 与 OfficeAce 的 MaaS 模型,实时模型与积分倍率、分桶积分用量和每日签到。取代 officeace-auth,原已登录的账号继续可用。" } }, - { "package": "@magpie-community/opencode-gcp-ai-auth", "name": "GCP AI", "providers": ["gcp-ai"], "community": true, - "summary": { "en": "Vertex AI models and Google Cloud promotional credits, signed with your own gcloud or ADC login. Lists only the models your project can call.", "zh": "用本机 gcloud / ADC 登录使用 Vertex AI 模型,并读取 Google Cloud 促销赠金;只列出项目实际能调用的模型。" } }, - { "package": "magpie-review-rules", "name": "Claude Desktop Review & Model Rules", "kind": "middleware", - "summary": {"en":"Pin Claude Desktop auto-mode action reviews to a chosen Magpie model, independently of the conversation model. Configure per-model Claude compatibility presets and sandboxed JavaScript request rules in a native settings page. No request changes until configured.","zh":"固定 Claude Desktop 自动模式的操作审核模型,不随对话模型切换;通过原生设置页配置逐模型 Claude 兼容预设与沙箱 JavaScript 请求规则。首次安装不改写请求。"} }, - { "package": "@magpie-community/middleware-param-override", "name": "Param Override", "kind": "middleware", "community": true, - "summary": { "en": "New API's parameter override: set, delete, move or rewrite fields of every request, by model or by condition.", "zh": "New API 的参数覆盖:按模型或条件设置、删除、移动或改写每个请求的字段。" } }, - { "package": "@magpie-community/middleware-model-map", "name": "Model Map", "kind": "middleware", "community": true, - "summary": { "en": "New API's model redirection: send a model under another name, by name or pattern; replies keep the name asked for.", "zh": "New API 的模型重定向:按名称或正则把模型换成另一个发出,回复里仍是原来的名字。" } }, - { "package": "@magpie-community/middleware-system-prompt", "name": "System Prompt", "kind": "middleware", "community": true, - "summary": { "en": "Add your own system prompt to every request, or only some agents' or models', in each API's own place.", "zh": "给每个请求(或指定 agent、模型)加上自己的系统提示词,放在各 API 自己的位置。" } }, - { "package": "@magpie-community/middleware-output-strategy", "name": "Output Strategy", "kind": "middleware", "community": true, - "summary": { "en": "Seven composable output styles, all off by default, with a selection panel and prompt-cache-stable injection that preserves Codex instructions. Unofficial; by manjughosa.", "zh": "七种可组合的输出样式,全部默认关闭;带选择面板,注入不破坏提示缓存,也不改写 Codex instructions。由 manjughosa 独立开发并维护,非官方出品。" } }, - { "package": "@magpie-community/middleware-word-guard", "name": "Word Guard", "kind": "middleware", "community": true, - "summary": { "en": "New API's sensitive-word filter: turn away or mask words and patterns in what you send, and in replies.", "zh": "New API 的敏感词过滤:拒绝或替换你发出的内容和回复里的词与正则。" } }, - { "package": "@magpie-community/middleware-company-guard", "name": "Company Guard", "kind": "middleware", "community": true, - "summary": { "en": "Recognize and mask company names with stable session aliases, restore replies, and show local receipts and explicit Codex session approvals. Unlike Word Guard, names can be restored. Text only; rule-based detection.", "zh": "识别公司名称并使用会话内稳定代号,回复返回本机时还原,提供本地回执与需明确确认的 Codex 会话授权。区别于 Word Guard 的词表拒绝或替换,支持名称还原;仅处理文字,采用规则识别。" } }, - { "package": "@magpie-community/middleware-think-tags", "name": "Think Tags", "kind": "middleware", "community": true, - "summary": { "en": "Take a model's leading , or block out of its reply's text, or put a Chat reply's reasoning into it.", "zh": "去掉回复正文开头的 、 或 思考块,或把 Chat 回复的思考内容放进正文。" } }, - { "package": "@magpie-community/middleware-empty-reply", "name": "Empty Reply", "kind": "middleware", "community": true, - "summary": { "en": "Drop the end of a reply that has reasoning but no text, so Claude Code and Codex retry instead of treating an empty turn as finished. Anthropic and Responses, streamed or whole.", "zh": "丢掉「只有思考、没有正文」回复的结尾,让 Claude Code 和 Codex 重试,而不是把空回合当成正常结束。支持 Anthropic 与 Responses 的流式与整包回复。" } }, - { "package": "@magpie-community/agent-nowledge-mem", "name": "Nowledge Mem / AI Now models", "kind": "agent", "community": true, "icon": "https://mem.nowledge.co/favicon.png", - "summary": {"en": "Choose Nowledge Mem Default, AI Now and Background models through Magpie. Distinct from gateway-nowledge-mem recall/capture. Disconnect before uninstalling.", "zh": "在 Magpie 管理 Nowledge 的 Default、AI Now 和后台模型;与 gateway-nowledge-mem 的记忆召回、对话导入用途不同。卸载前先断开。"} }, - { "package": "@magpie-community/agent-kilo", "name": "Kilo Code", "kind": "agent", "community": true, "icon": "https://github.com/Kilo-Org.png", - "summary": { "en": "Kilo Code (kilo.ai) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put its config back.", "zh": "把 Kilo Code(kilo.ai)加到 magpie 的 Agent 页:选择 magpie 的模型,关掉即恢复原来的配置。" } }, - { "package": "@magpie-community/middleware-antigravity-cloak", "name": "Antigravity Cloak", "kind": "middleware", "community": true, - "summary": { "en": "Claude Code and Claude Desktop on an Antigravity account: the words Antigravity refuses their system prompt for are broken up, in what Antigravity alone is sent.", "zh": "让 Claude Code 和 Claude Desktop 能用 Antigravity 账号:只在发给 Antigravity 的请求里拆开它据以拒绝系统提示词的那几个词。" } }, - { "package": "@magpie-community/middleware-fovea", "name": "Fovea", "kind": "middleware", "community": true, - "summary": { "en": "Thin an agent's old history while preserving a recent window and tool pairing. Optional saccade collapses old content on a new topic; a structured state block replaces the raw task message; cut boundaries move in blocks to keep provider prompt caches settling.", "zh": "裁薄 Agent 旧历史,保留最近窗口与工具配对;可选 saccade 在新话题时折叠旧内容,结构化状态块替代原始任务消息,裁剪边界分块移动以保护 prompt cache。" } }, - { "package": "@magpie-community/middleware-codex-history", "name": "Codex History Filter", "kind": "middleware", "community": true, - "summary": { "en": "Opt-in, per-provider Codex Responses history filtering after switching providers. Preserves readable content and tool-call pairing; no model probes.", "zh": "按供应商开关 Codex Responses 历史过滤,缓解切换上游后的密文和历史 ID 兼容问题;保留正文与工具配对,不运行模型检测。" } }, - { "package": "@magpie-community/middleware-history-trim", "name": "History Trim", "kind": "middleware", "community": true, - "summary": { "en": "Send a request's history smaller: fold a run of blank lines, drop a system prompt written twice and a message repeated right after itself, and cut a past tool result longer than maxToolLength at a word boundary. The turn being sent now is left whole.", "zh": "把请求的历史发得更小:折叠连续空行、去掉逐字重复的 system 提示与紧挨着重复的消息、把历史中超过 maxToolLength 的工具结果在词边界处截断;正在发送的这一轮原样不动。" } }, - { "package": "@magpie-community/agent-agy", "name": "Antigravity CLI", "kind": "agent", "community": true, "icon": "antigravity-color", - "summary": { "en": "Antigravity CLI (agy) on magpie's Agents page: pick one of magpie's models, and the agy in your shells starts on magpie's gateway. Takes the place of the built-in, keeping what it set up.", "zh": "把 Antigravity CLI(agy)加到 magpie 的 Agent 页:选一个 magpie 的模型,终端里的 agy 就从 magpie 的网关启动。接替内置的 agy,保留它原有的设置。" } }, - { "package": "@magpie-community/agent-aider", "name": "Aider", "kind": "agent", "community": true, - "summary": { "en": "Aider (aider.chat) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put its config back.", "zh": "把 Aider(aider.chat)加到 magpie 的 Agent 页:给它选 magpie 的模型,关掉即恢复它原来的配置。" } }, - { "package": "@magpie-community/agent-nanobot", "name": "nanobot", "kind": "agent", "community": true, "icon": "https://raw.githubusercontent.com/HKUDS/nanobot/main/webui/public/brand/nanobot_mark.svg", - "summary": { "en": "nanobot (HKUDS/nanobot) on magpie's Agents page: pick one of magpie's models for it, give it Library MCP servers and skills, and switch it off to put its config back.", "zh": "把 nanobot(HKUDS/nanobot)加到 magpie 的 Agent 页:给它选 magpie 的模型、勾选资源库的 MCP 服务与技能,关掉即恢复它原来的配置。" } }, - { "package": "@magpie-community/agent-reasonix-desktop", "name": "Reasonix 1.x", "kind": "agent", "community": true, "icon": "reasonix-color", - "summary": { "en": "Reasonix 1.x Desktop and CLI on Magpie's models, preserving existing providers. Restart Reasonix after choosing a model; disconnect before uninstalling. Separate from the built-in Reasonix Studio 2.x adapter.", "zh": "为 Reasonix 1.x 桌面版和 CLI 选择 Magpie 模型,保留原有供应商;选模型后重启 Reasonix,卸载前先断开。与内置的 Reasonix Studio 2.x 适配器分开。" } }, - { "package": "@magpie-community/agent-deepzora", "name": "Deepzora", "kind": "agent", "community": true, "icon": "https://raw.githubusercontent.com/SivanCola/deepzora/ad855397cdfdaf2ba7e3e3414583ba777412cbe4/desktop/build/linux/icons/hicolor/128x128/apps/deepzora-desktop.png", - "summary": { "en": "Deepzora Desktop, CLI and ACP on Magpie's models, with a reversible TOML/dotenv bridge that preserves existing providers. Restart Deepzora after choosing a model; disconnect before uninstalling. Disable the standalone Deepzora plugin before switching to this package.", "zh": "为 Deepzora 桌面版、CLI 和 ACP 选择 Magpie 模型,通过可恢复的 TOML/dotenv 桥接保留原有供应商。选模型后重启 Deepzora,卸载前先断开;从独立 Deepzora 插件迁移时先停用旧包。" } }, - { "package": "@magpie-community/agent-pig", "name": "PiG", "kind": "agent", "community": true, - "summary": { "en": "PiG (Pi in Go): add Magpie's endpoint and model catalog, choose models in PiG, and keep startup settings unchanged.", "zh": "为 PiG(Pi in Go)注入 Magpie 端点与模型列表,在 PiG 内选择模型,保持启动设置不变。" } }, - { "package": "@magpie-community/agent-bub", "name": "Bub", "kind": "agent", "community": true, - "summary": { "en": "Bub (bub.build) on magpie's Agents page: pick one of magpie's models and Bub uses it, through ~/.bub/config.yml. Needs Bub 0.5.1 or later.", "zh": "把 Bub(bub.build)加到 magpie 的 Agent 页:给它选一个 magpie 的模型,通过 ~/.bub/config.yml 生效。需要 Bub 0.5.1 或更新版本。" } }, - { "package": "@magpie-community/agent-type4me", "name": "Type4Me", "kind": "agent", "community": true, - "summary": { "en": "Type4Me (macOS voice input) on magpie's Agents page: pick one of magpie's models for its LLM post-processing, used from the next dictation.", "zh": "把 Type4Me(macOS 语音输入)加到 magpie 的 Agent 页:给它的 LLM 后处理选 magpie 的模型,下一次语音输入即生效。" } }, - { "package": "@magpie-community/agent-cowork", "name": "Tuanjie Cowork", "kind": "agent", "community": true, - "summary": { "en": "Tuanjie Cowork (codely) on magpie's Agents page: pick one of magpie's models and Cowork chats with it, on the subscriptions magpie already holds.", "zh": "把团结 Cowork(codely)加到 magpie 的 Agent 页:给它选一个 magpie 的模型,它聊天就走那个模型,用的是 magpie 里已登录的那些订阅的额度。" } }, - { "package": "@magpie-community/agent-opencodereview", "name": "Open Code Review", "kind": "agent", "community": true, - "summary": { "en": "Open Code Review (Alibaba's ocr CLI) on magpie's Agents page: pick one of magpie's models and ocr review / ocr scan use it, read from the config on the next run.", "zh": "把 Open Code Review(阿里巴巴的 ocr CLI)加到 magpie 的 Agent 页:给它选 magpie 的模型,下一次 ocr review / ocr scan 即从配置中读到并生效。" } }, - { "package": "@magpie-community/agent-pi-desktop", "name": "Pi-Desktop", "kind": "agent", "community": true, - "summary": { "en": "Pi-Desktop on magpie's Agents page: magpie keeps a Pi-Desktop plugin folder that adds it as a model provider, with its current models and their thinking levels. Pi-Desktop picks up model changes when it starts.", "zh": "把 Pi-Desktop 加到 magpie 的 Agent 页:magpie 维护一个 Pi-Desktop 插件文件夹,把自己加为模型服务,带上当前模型和推理等级;模型有变化时,Pi-Desktop 重启后生效。" } }, - { "package": "@magpie-community/agent-junie", "name": "Junie CLI", "kind": "agent", "community": true, + "summary": { + "en": "Recall local session history from the transcripts Claude Code, Codex, Hermes, Gemini CLI, OpenCode, Cursor and DSH write on this machine before each request. Scopes lookup to the request's project directory (ctx.cwd) by default and scrubs secret patterns before attaching recalled excerpts to the request sent to your model provider. When recalled history helped, the model is asked to end its reply with a one-line \"déjà vu: … — reusing it.\" note.", + "zh": "每次请求前从本机 Claude Code、Codex、Hermes、Gemini CLI、OpenCode、Cursor 与 DSH 的磁盘会话记录中召回相关历史。默认严格限定在请求所属的项目目录(ctx.cwd)内检索,并对密钥与令牌执行脱敏后,将召回片段随请求一并发送给所配置的模型供应商。当召回的历史记录有帮助时,会请模型在回复末尾附上一行“déjà vu: … — reusing it.”说明。" + } + }, + { + "package": "@magpie-community/opencode-copilot-auth", + "name": "GitHub Copilot", + "icon": "githubcopilot", + "providers": [ + "github-copilot" + ], + "community": true, + "summary": { + "en": "GitHub Copilot on GitHub.com or GitHub Enterprise, with premium request or credit usage, remaining allowance and quota reset dates.", + "zh": "GitHub Copilot:支持 GitHub.com 与 GitHub Enterprise,显示高级请求或积分用量、剩余额度和重置时间。" + } + }, + { + "package": "@magpie-community/opencode-joycode-auth", + "name": "JoyCode", + "providers": [ + "joycode" + ], + "community": true, + "summary": { + "en": "Unofficial JoyCode provider: JD account browser sign-in, live models and credit multipliers, plan credits and streaming tool calls.", + "zh": "非官方 JoyCode 插件:京东账号浏览器登录,动态模型与积分倍率、套餐额度及流式工具调用。" + } + }, + { + "package": "@magpie-community/opencode-doubao-work-auth", + "name": "豆包工作", + "icon": "doubao-color", + "providers": [ + "doubao-work" + ], + "community": true, + "summary": { + "en": "Not a model API: Doubao Work has none for other clients. This drives its desktop app (or Doubao's) the way a person does, so each request is one run of Doubao's cloud agent in a new task, with the client's tools asked for as text. macOS and Windows; unofficial.", + "zh": "不是模型接口:豆包工作没有给第三方客户端用的模型接口。本插件像人一样操作豆包工作(或豆包)桌面版,每个请求是豆包云端智能体在一个新任务里的一次执行,客户端的工具以文本方式调用。macOS 和 Windows;非官方。" + } + }, + { + "package": "@magpie-community/opencode-kimi-work-auth", + "name": "Kimi Work", + "icon": "kimi", + "providers": [ + "kimi-work" + ], + "community": true, + "summary": { + "en": "Kimi desktop Work subscription through its local sign-in, with live Work models and shared membership usage. Separate from Kimi Code.", + "zh": "使用本机普通 Kimi 桌面版的 Work 登录,自动读取 Work 模型和会员共享用量;独立于 Kimi Code。" + } + }, + { + "package": "@magpie-community/opencode-zen-free-auth", + "name": "OpenCode Zen Free", + "icon": "opencode", + "providers": [ + "opencode-zen-free" + ], + "community": true, + "summary": { + "en": "OpenCode Zen's free models with the public credential, automatic discovery and native streaming protocols.", + "zh": "使用 public 凭证接入 OpenCode Zen 免费模型,自动发现模型并支持原生流式协议。" + } + }, + { + "package": "@magpie-community/opencode-zcode-auth", + "name": "ZCode", + "icon": "zcode", + "providers": [ + "zcode" + ], + "community": true, + "replaces": "zcode", + "summary": { + "en": "Z.ai and BigModel GLM Coding Plan, team seats and ZCode's free Start Plan, with a daily claim of its gift plans.", + "zh": "Z.ai 与智谱 GLM Coding Plan,支持团队席位和 ZCode 免费 Start Plan,可每日领取 ZCode 活动包。" + } + }, + { + "package": "@magpie-community/opencode-qoder-auth", + "name": "Qoder", + "icon": "qoder", + "providers": [ + "qoder", + "qoder-cn", + "qwenwork", + "qwenwork-enterprise" + ], + "community": true, + "replaces": "qoder", + "summary": { + "en": "Your Qoder subscription: international (qoder.com) or Qoder CN, the China site (qoder.cn, Alibaba Cloud or phone sign-in), each its own provider. Also QwenWork (千问办公): the public cloud or the enterprise flagship with its own domain. Unofficial for QwenWork.", + "zh": "用 Qoder 自己的设备登录使用 Qoder 订阅:国际版(qoder.com)与 Qoder 中国版(Qoder CN,qoder.cn,阿里云账号 / 手机号登录)各是一个订阅;也接入千问办公:公有云与企业旗舰版(输入企业自己的域名)。千问办公部分非官方。" + } + }, + { + "package": "@magpie-community/opencode-devin-auth", + "name": "Devin", + "icon": "devin", + "providers": [ + "devin" + ], + "community": true, + "replaces": "devin", + "summary": { + "en": "Devin's models with your Devin account — a browser sign-in or the devin CLI's.", + "zh": "用 Devin 账号使用 Devin 的模型:浏览器登录,或沿用 devin CLI 的登录。" + } + }, + { + "package": "@magpie-community/opencode-zed-auth", + "name": "Zed", + "icon": "zed", + "providers": [ + "zed" + ], + "community": true, + "replaces": "zed", + "summary": { + "en": "Claude, GPT, Gemini and Grok hosted by Zed, on your Zed plan.", + "zh": "用 Zed 订阅使用 Zed 托管的 Claude、GPT、Gemini 和 Grok。" + } + }, + { + "package": "@magpie-community/opencode-factory-auth", + "name": "Factory", + "icon": "factory", + "providers": [ + "factory" + ], + "community": true, + "replaces": "factory", + "summary": { + "en": "Factory's Droid subscription: models from Anthropic, OpenAI, xAI and open labs.", + "zh": "Factory Droid 订阅:Anthropic、OpenAI、xAI 及开源模型。" + } + }, + { + "package": "@magpie-community/opencode-grok-auth", + "name": "Grok", + "icon": "xai", + "providers": [ + "grok" + ], + "community": true, + "replaces": "grok", + "summary": { + "en": "SuperGrok and X Premium+, through the Grok Build CLI's sign-in.", + "zh": "通过 Grok Build CLI 的登录使用 SuperGrok 和 X Premium+。" + } + }, + { + "package": "@magpie-community/opencode-commandcode-auth", + "name": "Command Code", + "icon": "commandcode", + "providers": [ + "commandcode-plan" + ], + "community": true, + "replaces": "commandcode-plan", + "summary": { + "en": "Command Code plans — Pro, GOAT, Max, Ultra, Go and Teams.", + "zh": "Command Code 各档套餐:Pro、GOAT、Max、Ultra、Go 和 Teams。" + } + }, + { + "package": "@magpie-community/opencode-mimo-auth", + "name": "Xiaomi MiMo", + "icon": "mimocode", + "providers": [ + "mimo-app" + ], + "community": true, + "replaces": "mimo-app", + "summary": { + "en": "Xiaomi MiMo models with your Xiaomi account.", + "zh": "用小米账号使用小米 MiMo 模型。" + } + }, + { + "package": "@magpie-community/opencode-cursor-auth", + "name": "Cursor", + "icon": "cursor", + "providers": [ + "cursor" + ], + "community": true, + "replaces": "cursor", + "summary": { + "en": "Cursor Pro, Pro+, Ultra and Teams — every model your plan has, with tool calls.", + "zh": "Cursor Pro、Pro+、Ultra 与 Teams:套餐里的所有模型,支持工具调用。" + } + }, + { + "package": "@magpie-community/opencode-kiro-auth", + "name": "Kiro", + "icon": "kiro-color", + "providers": [ + "kiro" + ], + "community": true, + "replaces": "kiro", + "summary": { + "en": "Kiro Free, Pro, Pro+ and Power — Kiro's sign-in, kiro-cli's, an API key, or accounts imported by refresh token.", + "zh": "Kiro Free、Pro、Pro+ 与 Power:Kiro 登录、沿用 kiro-cli 的登录、API key,或按 refresh token 导入账号。" + } + }, + { + "package": "@magpie-community/opencode-minimax-auth", + "name": "MiniMax Code", + "icon": "minimax-color", + "providers": [ + "minimax-code", + "minimax-code-global" + ], + "community": true, + "summary": { + "en": "MiniMax Code credits and M Plan with device sign-in, native web search, image/video generation and media tools through Magpie MCP.", + "zh": "MiniMax Code 积分与 M Plan、设备登录,提供原生网络搜索、生图、生视频及多模态 MCP 工具。" + } + }, + { + "package": "@magpie-community/opencode-warp-auth", + "name": "Warp", + "icon": "warp", + "providers": [ + "warp" + ], + "community": true, + "summary": { + "en": "Warp (the terminal) AI request credits through the Warp app's sign-in: the auto routers and every model family the account lists.", + "zh": "用 Warp 终端应用的登录使用 Warp AI 请求额度:auto 路由与账户可用的全部模型家族。" + } + }, + { + "package": "@magpie-community/opencode-workbuddy-auth", + "name": "WorkBuddy", + "icon": "workbuddy-color", + "providers": [ + "workbuddy", + "workbuddy-ai" + ], + "community": true, + "replaces": "workbuddy", + "summary": { + "en": "WorkBuddy's China build (CodeBuddy plan) and its international build.", + "zh": "WorkBuddy 国内版(CodeBuddy 套餐)与国际版。" + } + }, + { + "package": "@magpie-community/opencode-cline-auth", + "name": "Cline", + "icon": "cline", + "providers": [ + "cline" + ], + "community": true, + "summary": { + "en": "Cline's models on your own Cline account — usage billing, ClinePass and Cline's free models — with Cline's device sign-in or an API key.", + "zh": "用自己的 Cline 账号使用 Cline 的模型——用量计费、ClinePass 与免费模型——支持 Cline 设备登录或 API key。" + } + }, + { + "package": "@magpie-community/opencode-trae-auth", + "name": "Trae", + "icon": "trae", + "providers": [ + "trae-cn", + "trae-global" + ], + "community": true, + "summary": { + "en": "Trae CN (trae.cn) and Trae international (trae.ai) models with your own account, free tier included — Trae's own browser sign-in. Experimental.", + "zh": "用 Trae CN(trae.cn)与 Trae 国际版(trae.ai)账号使用各自的模型,含免费版,使用 Trae 自己的浏览器登录。实验性。" + } + }, + { + "package": "@magpie-community/opencode-opendesign-auth", + "name": "Open Design", + "providers": [ + "opendesign" + ], + "community": true, + "summary": { + "en": "Open Design Cloud's Design Plan (open-design.ai): the account's device sign-in, the models the plan entitles it to, and the plan's 5-hour and 7-day allowance and wallet balance.", + "zh": "Open Design Cloud 的 Design Plan(open-design.ai):账号设备登录、套餐可用的模型,以及 5 小时与 7 天额度、钱包余额。" + } + }, + { + "package": "@magpie-community/opencode-catpaw-auth", + "name": "CatPaw", + "providers": [ + "catpaw" + ], + "community": true, + "summary": { + "en": "Meituan's 妙手 CatPaw (ai.catpaw.meituan.com): a browser web sign-in or the desktop client's saved credentials, chat through its round/event/turn conversation protocol, models and credit balance.", + "zh": "美团妙手 CatPaw(ai.catpaw.meituan.com):浏览器网页登录或妙手桌面端已保存的登录态,经其 round/event/turn 会话协议对话,含模型与积分余额。" + } + }, + { + "package": "@magpie-community/opencode-loomy-auth", + "name": "Loomy", + "providers": [ + "loomy" + ], + "community": true, + "summary": { + "en": "iFlytek's Loomy (loomyad.xunfei.cn): SMS sign-in, OpenAI-compatible chat with reasoning-effort folding and image markers, the live model catalog, two point wallets and the daily first-login grant.", + "zh": "讯飞 Loomy(loomyad.xunfei.cn):短信验证码登录,OpenAI 兼容对话(思考档折叠与图片标记适配),实时模型目录、双积分钱包与每日首登赠送。" + } + }, + { + "package": "@magpie-community/opencode-nous-auth", + "name": "Nous Portal (OAuth)", + "icon": "nousresearch", + "providers": [ + "nous-portal" + ], + "community": true, + "summary": { + "en": "Nous Portal models, with Nous Portal's device sign-in.", + "zh": "通过 Nous Portal 自己的设备码登录使用 Nous Portal 模型。" + } + }, + { + "package": "@magpie-community/opencode-raccoon-auth", + "name": "商汤小浣熊", + "icon": "sensenova-color", + "providers": [ + "raccoon" + ], + "community": true, + "summary": { + "en": "SenseTime Raccoon (xiaohuanxiong.com) models, account credits and daily check-in, with browser sign-in or Raccoon desktop's live sign-in.", + "zh": "商汤小浣熊(xiaohuanxiong.com)模型、积分余额与每日签到,支持浏览器登录或直接复用小浣熊桌面端登录态。" + } + }, + { + "package": "@magpie-community/opencode-sensenova-auth", + "name": "SenseNova", + "icon": "sensenova-color", + "providers": [ + "sensenova-token-plan" + ], + "community": true, + "summary": { + "en": "SenseNova (SenseTime 日日新) Token Plan: a browser sign-in (or the console account and password) reads the credit pools, their 5-hour and 7-day windows and the expiring grants, and serves the plan's own models — the browser sign-in reuses or creates a Token Plan API key for you.", + "zh": "商汤日日新 Token Plan:浏览器登录(或控制台账号密码)读取积分池、5 小时 / 7 天窗口与将过期的返赠余额,并直接调用计划内模型——浏览器登录会自动复用或创建 Token Plan API Key。" + } + }, + { + "package": "@magpie-community/opencode-inkstone-auth", + "name": "书生·端砚", + "icon": "https://raw.githubusercontent.com/magpie-community/plugins/main/packages/inkstone/icon.svg", + "providers": [ + "intern-inkstone" + ], + "community": true, + "summary": { + "en": "Unofficial: Intern InkStone(书生·端砚) Token Plan: a browser sign-in reads the ink-credit balance, the 5-hour and 7-day windows and the credit packs, and creates an ink-* API key (or reuses a pasted one) to serve the plan's models for chat.", + "zh": "非官方:书生·端砚(Intern InkStone)墨点计划:浏览器登录读取墨点余额、5 小时 / 7 天窗口与墨点包,并创建 ink-* API Key(或复用现有 Key)直接调用计划内模型对话。" + } + }, + { + "package": "@magpie-community/opencode-autoclaw-auth", + "name": "AutoClaw", + "providers": [ + "autoclaw", + "autoclaw-cn" + ], + "community": true, + "summary": { + "en": "AutoClaw (Zhipu OpenClaw) cloud accounts: oversea web sign-in (Google/z.ai) or CN phone SMS, credit balance and daily check-in.", + "zh": "AutoClaw(智谱澳龙/OpenClaw)云端账号:国际版网页登录(Google/z.ai)或国内手机验证码,积分余额与每日签到。" + } + }, + { + "package": "@magpie-community/opencode-codearts-auth", + "name": "CodeArts / OfficeAce", + "providers": [ + "codearts", + "officeace" + ], + "community": true, + "summary": { + "en": "Unofficial: Huawei Cloud CodeArts (码道) coding plans and OfficeAce (AgentArts), each signed in with Huawei Cloud OAuth on its own card — GLM, DeepSeek, openPangu and OfficeAce's MaaS models, live models and credit rates, credit usage by bucket and the daily check-in. Replaces officeace-auth; accounts signed in there keep working.", + "zh": "非官方:华为云码道(CodeArts)编码套餐与 OfficeAce(AgentArts),两张卡各自用华为云 OAuth 登录;GLM、DeepSeek、openPangu 与 OfficeAce 的 MaaS 模型,实时模型与积分倍率、分桶积分用量和每日签到。取代 officeace-auth,原已登录的账号继续可用。" + } + }, + { + "package": "@magpie-community/opencode-gcp-ai-auth", + "name": "GCP AI", + "providers": [ + "gcp-ai" + ], + "community": true, + "summary": { + "en": "Vertex AI models and Google Cloud promotional credits, signed with your own gcloud or ADC login. Lists only the models your project can call.", + "zh": "用本机 gcloud / ADC 登录使用 Vertex AI 模型,并读取 Google Cloud 促销赠金;只列出项目实际能调用的模型。" + } + }, + { + "package": "magpie-review-rules", + "name": "Claude Desktop Review & Model Rules", + "kind": "middleware", + "summary": { + "en": "Pin Claude Desktop auto-mode action reviews to a chosen Magpie model, independently of the conversation model. Configure per-model Claude compatibility presets and sandboxed JavaScript request rules in a native settings page. No request changes until configured.", + "zh": "固定 Claude Desktop 自动模式的操作审核模型,不随对话模型切换;通过原生设置页配置逐模型 Claude 兼容预设与沙箱 JavaScript 请求规则。首次安装不改写请求。" + } + }, + { + "package": "@magpie-community/middleware-param-override", + "name": "Param Override", + "kind": "middleware", + "community": true, + "summary": { + "en": "New API's parameter override: set, delete, move or rewrite fields of every request, by model or by condition.", + "zh": "New API 的参数覆盖:按模型或条件设置、删除、移动或改写每个请求的字段。" + } + }, + { + "package": "@magpie-community/middleware-model-map", + "name": "Model Map", + "kind": "middleware", + "community": true, + "summary": { + "en": "New API's model redirection: send a model under another name, by name or pattern; replies keep the name asked for.", + "zh": "New API 的模型重定向:按名称或正则把模型换成另一个发出,回复里仍是原来的名字。" + } + }, + { + "package": "@magpie-community/middleware-system-prompt", + "name": "System Prompt", + "kind": "middleware", + "community": true, + "summary": { + "en": "Add your own system prompt to every request, or only some agents' or models', in each API's own place.", + "zh": "给每个请求(或指定 agent、模型)加上自己的系统提示词,放在各 API 自己的位置。" + } + }, + { + "package": "@magpie-community/middleware-output-strategy", + "name": "Output Strategy", + "kind": "middleware", + "community": true, + "summary": { + "en": "Seven composable output styles, all off by default, with a selection panel and prompt-cache-stable injection that preserves Codex instructions. Unofficial; by manjughosa.", + "zh": "七种可组合的输出样式,全部默认关闭;带选择面板,注入不破坏提示缓存,也不改写 Codex instructions。由 manjughosa 独立开发并维护,非官方出品。" + } + }, + { + "package": "@magpie-community/middleware-word-guard", + "name": "Word Guard", + "kind": "middleware", + "community": true, + "summary": { + "en": "New API's sensitive-word filter: turn away or mask words and patterns in what you send, and in replies.", + "zh": "New API 的敏感词过滤:拒绝或替换你发出的内容和回复里的词与正则。" + } + }, + { + "package": "@magpie-community/middleware-company-guard", + "name": "Company Guard", + "kind": "middleware", + "community": true, + "summary": { + "en": "Recognize and mask company names with stable session aliases, restore replies, and show local receipts and explicit Codex session approvals. Unlike Word Guard, names can be restored. Text only; rule-based detection.", + "zh": "识别公司名称并使用会话内稳定代号,回复返回本机时还原,提供本地回执与需明确确认的 Codex 会话授权。区别于 Word Guard 的词表拒绝或替换,支持名称还原;仅处理文字,采用规则识别。" + } + }, + { + "package": "@magpie-community/middleware-think-tags", + "name": "Think Tags", + "kind": "middleware", + "community": true, + "summary": { + "en": "Take a model's leading , or block out of its reply's text, or put a Chat reply's reasoning into it.", + "zh": "去掉回复正文开头的 、 或 思考块,或把 Chat 回复的思考内容放进正文。" + } + }, + { + "package": "@magpie-community/middleware-empty-reply", + "name": "Empty Reply", + "kind": "middleware", + "community": true, + "summary": { + "en": "Drop the end of a reply that has reasoning but no text, so Claude Code and Codex retry instead of treating an empty turn as finished. Anthropic and Responses, streamed or whole.", + "zh": "丢掉「只有思考、没有正文」回复的结尾,让 Claude Code 和 Codex 重试,而不是把空回合当成正常结束。支持 Anthropic 与 Responses 的流式与整包回复。" + } + }, + { + "package": "@magpie-community/agent-nowledge-mem", + "name": "Nowledge Mem / AI Now models", + "kind": "agent", + "community": true, + "icon": "https://mem.nowledge.co/favicon.png", + "summary": { + "en": "Choose Nowledge Mem Default, AI Now and Background models through Magpie. Distinct from gateway-nowledge-mem recall/capture. Disconnect before uninstalling.", + "zh": "在 Magpie 管理 Nowledge 的 Default、AI Now 和后台模型;与 gateway-nowledge-mem 的记忆召回、对话导入用途不同。卸载前先断开。" + } + }, + { + "package": "@magpie-community/agent-kilo", + "name": "Kilo Code", + "kind": "agent", + "community": true, + "icon": "https://github.com/Kilo-Org.png", + "summary": { + "en": "Kilo Code (kilo.ai) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put its config back.", + "zh": "把 Kilo Code(kilo.ai)加到 magpie 的 Agent 页:选择 magpie 的模型,关掉即恢复原来的配置。" + } + }, + { + "package": "@magpie-community/middleware-antigravity-cloak", + "name": "Antigravity Cloak", + "kind": "middleware", + "community": true, + "summary": { + "en": "Claude Code and Claude Desktop on an Antigravity account: the words Antigravity refuses their system prompt for are broken up, in what Antigravity alone is sent.", + "zh": "让 Claude Code 和 Claude Desktop 能用 Antigravity 账号:只在发给 Antigravity 的请求里拆开它据以拒绝系统提示词的那几个词。" + } + }, + { + "package": "@magpie-community/middleware-fovea", + "name": "Fovea", + "kind": "middleware", + "community": true, + "summary": { + "en": "Thin an agent's old history while preserving a recent window and tool pairing. Optional saccade collapses old content on a new topic; a structured state block replaces the raw task message; cut boundaries move in blocks to keep provider prompt caches settling.", + "zh": "裁薄 Agent 旧历史,保留最近窗口与工具配对;可选 saccade 在新话题时折叠旧内容,结构化状态块替代原始任务消息,裁剪边界分块移动以保护 prompt cache。" + } + }, + { + "package": "@magpie-community/middleware-codex-history", + "name": "Codex History Filter", + "kind": "middleware", + "community": true, + "summary": { + "en": "Opt-in, per-provider Codex Responses history filtering after switching providers. Preserves readable content and tool-call pairing; no model probes.", + "zh": "按供应商开关 Codex Responses 历史过滤,缓解切换上游后的密文和历史 ID 兼容问题;保留正文与工具配对,不运行模型检测。" + } + }, + { + "package": "@magpie-community/middleware-history-trim", + "name": "History Trim", + "kind": "middleware", + "community": true, + "summary": { + "en": "Send a request's history smaller: fold a run of blank lines, drop a system prompt written twice and a message repeated right after itself, and cut a past tool result longer than maxToolLength at a word boundary. The turn being sent now is left whole.", + "zh": "把请求的历史发得更小:折叠连续空行、去掉逐字重复的 system 提示与紧挨着重复的消息、把历史中超过 maxToolLength 的工具结果在词边界处截断;正在发送的这一轮原样不动。" + } + }, + { + "package": "@magpie-community/agent-agy", + "name": "Antigravity CLI", + "kind": "agent", + "community": true, + "icon": "antigravity-color", + "summary": { + "en": "Antigravity CLI (agy) on magpie's Agents page: pick one of magpie's models, and the agy in your shells starts on magpie's gateway. Takes the place of the built-in, keeping what it set up.", + "zh": "把 Antigravity CLI(agy)加到 magpie 的 Agent 页:选一个 magpie 的模型,终端里的 agy 就从 magpie 的网关启动。接替内置的 agy,保留它原有的设置。" + } + }, + { + "package": "@magpie-community/agent-aider", + "name": "Aider", + "kind": "agent", + "community": true, + "summary": { + "en": "Aider (aider.chat) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put its config back.", + "zh": "把 Aider(aider.chat)加到 magpie 的 Agent 页:给它选 magpie 的模型,关掉即恢复它原来的配置。" + } + }, + { + "package": "@magpie-community/agent-nanobot", + "name": "nanobot", + "kind": "agent", + "community": true, + "icon": "https://raw.githubusercontent.com/HKUDS/nanobot/main/webui/public/brand/nanobot_mark.svg", + "summary": { + "en": "nanobot (HKUDS/nanobot) on magpie's Agents page: pick one of magpie's models for it, give it Library MCP servers and skills, and switch it off to put its config back.", + "zh": "把 nanobot(HKUDS/nanobot)加到 magpie 的 Agent 页:给它选 magpie 的模型、勾选资源库的 MCP 服务与技能,关掉即恢复它原来的配置。" + } + }, + { + "package": "@magpie-community/agent-reasonix-desktop", + "name": "Reasonix 1.x", + "kind": "agent", + "community": true, + "icon": "reasonix-color", + "summary": { + "en": "Reasonix 1.x Desktop and CLI on Magpie's models, preserving existing providers. Restart Reasonix after choosing a model; disconnect before uninstalling. Separate from the built-in Reasonix Studio 2.x adapter.", + "zh": "为 Reasonix 1.x 桌面版和 CLI 选择 Magpie 模型,保留原有供应商;选模型后重启 Reasonix,卸载前先断开。与内置的 Reasonix Studio 2.x 适配器分开。" + } + }, + { + "package": "@magpie-community/agent-deepzora", + "name": "Deepzora", + "kind": "agent", + "community": true, + "icon": "https://raw.githubusercontent.com/SivanCola/deepzora/ad855397cdfdaf2ba7e3e3414583ba777412cbe4/desktop/build/linux/icons/hicolor/128x128/apps/deepzora-desktop.png", + "summary": { + "en": "Deepzora Desktop, CLI and ACP on Magpie's models, with a reversible TOML/dotenv bridge that preserves existing providers. Restart Deepzora after choosing a model; disconnect before uninstalling. Disable the standalone Deepzora plugin before switching to this package.", + "zh": "为 Deepzora 桌面版、CLI 和 ACP 选择 Magpie 模型,通过可恢复的 TOML/dotenv 桥接保留原有供应商。选模型后重启 Deepzora,卸载前先断开;从独立 Deepzora 插件迁移时先停用旧包。" + } + }, + { + "package": "@magpie-community/agent-pig", + "name": "PiG", + "kind": "agent", + "community": true, + "summary": { + "en": "PiG (Pi in Go): add Magpie's endpoint and model catalog, choose models in PiG, and keep startup settings unchanged.", + "zh": "为 PiG(Pi in Go)注入 Magpie 端点与模型列表,在 PiG 内选择模型,保持启动设置不变。" + } + }, + { + "package": "@magpie-community/agent-bub", + "name": "Bub", + "kind": "agent", + "community": true, + "summary": { + "en": "Bub (bub.build) on magpie's Agents page: pick one of magpie's models and Bub uses it, through ~/.bub/config.yml. Needs Bub 0.5.1 or later.", + "zh": "把 Bub(bub.build)加到 magpie 的 Agent 页:给它选一个 magpie 的模型,通过 ~/.bub/config.yml 生效。需要 Bub 0.5.1 或更新版本。" + } + }, + { + "package": "@magpie-community/agent-type4me", + "name": "Type4Me", + "kind": "agent", + "community": true, + "summary": { + "en": "Type4Me (macOS voice input) on magpie's Agents page: pick one of magpie's models for its LLM post-processing, used from the next dictation.", + "zh": "把 Type4Me(macOS 语音输入)加到 magpie 的 Agent 页:给它的 LLM 后处理选 magpie 的模型,下一次语音输入即生效。" + } + }, + { + "package": "@magpie-community/agent-cowork", + "name": "Tuanjie Cowork", + "kind": "agent", + "community": true, + "summary": { + "en": "Tuanjie Cowork (codely) on magpie's Agents page: pick one of magpie's models and Cowork chats with it, on the subscriptions magpie already holds.", + "zh": "把团结 Cowork(codely)加到 magpie 的 Agent 页:给它选一个 magpie 的模型,它聊天就走那个模型,用的是 magpie 里已登录的那些订阅的额度。" + } + }, + { + "package": "@magpie-community/agent-opencodereview", + "name": "Open Code Review", + "kind": "agent", + "community": true, + "summary": { + "en": "Open Code Review (Alibaba's ocr CLI) on magpie's Agents page: pick one of magpie's models and ocr review / ocr scan use it, read from the config on the next run.", + "zh": "把 Open Code Review(阿里巴巴的 ocr CLI)加到 magpie 的 Agent 页:给它选 magpie 的模型,下一次 ocr review / ocr scan 即从配置中读到并生效。" + } + }, + { + "package": "@magpie-community/agent-pi-desktop", + "name": "Pi-Desktop", + "kind": "agent", + "community": true, + "summary": { + "en": "Pi-Desktop on magpie's Agents page: magpie keeps a Pi-Desktop plugin folder that adds it as a model provider, with its current models and their thinking levels. Pi-Desktop picks up model changes when it starts.", + "zh": "把 Pi-Desktop 加到 magpie 的 Agent 页:magpie 维护一个 Pi-Desktop 插件文件夹,把自己加为模型服务,带上当前模型和推理等级;模型有变化时,Pi-Desktop 重启后生效。" + } + }, + { + "package": "@magpie-community/agent-junie", + "name": "Junie CLI", + "kind": "agent", + "community": true, "icon": "https://junie.jetbrains.com/favicons/favicon.svg", - "summary": { "en": "An explicit seed copies Magpie's full model catalog into Junie CLI and uses Magpie's selected model as the launch default, without LiteLLM. Native ECO is unavailable for custom profiles in Junie 3893.6.", "zh": "手动 seed 向 Junie CLI 同步全部模型,并将 Magpie 所选模型设为启动默认模型,无需 LiteLLM。Junie 3893.6 的自定义 profile 不支持原生 ECO。" } }, - { "package": "@magpie-community/middleware-reserve", "name": "Reserve", "kind": "middleware", "community": true, - "summary": { "en": "Two routings that keep something in reserve: fill accounts up to a ceiling of their allowance, or use subscriptions before API keys.", "zh": "两种留余量的路由:账号用到额度上限就换下一个,或订阅账号优先于 API key。" } }, - { "package": "@magpie-community/tools-search1api", "name": "Search1API", "icon": "https://s1.dev/icon.png", "kind": "tools", "community": true, - "summary": { "en": "Search1API web search, news, agentic search, page retrieval, sitemaps and trends through Magpie's tools MCP server. Sign in with OAuth or an API key.", "zh": "通过 Magpie 的工具 MCP 服务使用 Search1API 网页搜索、新闻、智能搜索、页面读取、站点链接和趋势;支持 OAuth 或 API key 登录。" } }, - { "package": "@guotree/magpie-websearch", "name": "magpie-websearch", "icon": "https://raw.githubusercontent.com/guotree/magpie-websearch/main/assets/icon-128.png", "kind": "tools", "community": true, - "summary": { "en": "Adds 19 search APIs to Magpie: AnySearch, Bocha, Jina, Kagi, Keenable, Linkup, Metaso, Moonshot, Moonshot CN, Ollama, Parallel, Perplexity, Querit, SerpApi, Serper, StepFun, TinyFish, Z.ai, Zhipu. 4 support keyless access; 15 require keys.", "zh": "为 Magpie 网页搜索增加 19 家搜索源:AnySearch、Bocha、Jina、Kagi、Keenable、Linkup、Metaso、Moonshot、Moonshot CN、Ollama、Parallel、Perplexity、Querit、SerpApi、Serper、StepFun、TinyFish、Z.ai 和智谱。4 家支持免 Key,其余 15 家需要 API Key。" } }, - { "package": "magpie-search-mix", "name": "magpie-search-mix", "kind": "tools", "community": true, - "summary": { "en": "Mixes two search APIs for the AI: more accurate than one. Try DeepSeek + Tavily or You.com.", "zh": "混合搜索:同时问两家搜索 API,比只用一家更准。如 DeepSeek 搭配 Tavily 或 You.com。" } }, - { "package": "magpie-comfyui-link", "name": "ComfyUI", "kind": "tools", "community": true, - "summary": { "en": "A bridge between Magpie agents and locally hosted ComfyUI models, exposing native workflows, model resources and generated images through Magpie's tools MCP endpoint.", "zh": "连接 Magpie Agent 与本地 ComfyUI 模型的桥接层,通过 Magpie 工具 MCP 入口提供原生工作流、模型资源和生成图片。" } }, - { "package": "@magpie-community/tools-specstream", "name": "SpecStream", "kind": "tools", "community": true, - "summary": { "en": "SpecStream web search and page extraction for every agent: answers with the source page of each value, and the text of up to 20 URLs a call. Needs a SpecStream API key.", "zh": "给每个 agent 的 SpecStream 网页搜索和网页提取:每个值都带来源页面,一次最多读取 20 个网址的正文。需要 SpecStream API key。" } }, - { "package": "@magpie-community/tools-calc", "name": "Calc", "kind": "tools", "community": true, - "summary": { "en": "Exact arithmetic, unit conversion and calendar date math as tools for every agent: no network, no key.", "zh": "给每个 agent 的精确计算、单位换算和日历日期计算工具:不联网,不需要 key。" } }, - {"package": "@magpie-community/tools-layerfs", "name": "LayerFS", "kind": "tools", "community": true, "summary": {"en": "LayerFS workspaces for every agent: deduplicated checkpoints and branches behind stable filesystem paths, instead of a fresh copy of the tree for every agent.", "zh": "给每个 agent 的 LayerFS 工作区:稳定的文件系统路径背后是可去重的检查点与分支,而不是每个 agent 各复制一棵项目树。"}}, - { "package": "@magpie-community/page-usage", "name": "Usage Glance", "kind": "page", "community": true, - "summary": { "en": "A page in magpie's window with calls, errors, tokens and cost for today, 7 or 30 days, and the models behind them. Reads only your usage summary.", "zh": "在 magpie 窗口里一眼看到今天、7 天或 30 天的调用、错误、token、费用和对应模型。只读取用量汇总。" } }, - {"package": "@magpie-community/page-amp-usage", "name": "Amp Usage", "kind": "page", "community": true, "icon": "https://raw.githubusercontent.com/magpie-community/plugins/main/packages/amp-usage/page/icon.svg", "summary": {"en": "An Amp usage board: requests, tokens and cache hit rate by hour, day and model, read from your Amp account and legacy local threads. Its background service syncs that usage into Magpie requests and Day card, with duplicate checks; set sync to false to only view it.", "zh": "Amp \u7528\u91cf\u770b\u677f\uff1a\u8bf7\u6c42\u6570\u3001Token\u3001\u7f13\u5b58\u547d\u4e2d\u7387\uff0c\u6309\u5c0f\u65f6\u3001\u6309\u65e5\u548c\u6309\u6a21\u578b\u67e5\u770b\uff0c\u6570\u636e\u6765\u81ea Amp \u8d26\u53f7\u548c\u65e7\u7248\u672c\u5730\u7ebf\u7a0b\u3002\u540e\u53f0\u670d\u52a1\u628a\u8fd9\u4e9b\u7528\u91cf\u540c\u6b65\u8fdb Magpie \u8bf7\u6c42\u7edf\u8ba1\u548c\u6bcf\u65e5\u5361\u7247\uff0c\u5e26\u53bb\u91cd\u68c0\u67e5\uff1b\u628a sync \u8bbe\u4e3a false \u5219\u53ea\u67e5\u770b\u3002"}}, - { "package": "@magpie-community/page-day-card", "name": "Day Card", "kind": "page", "community": true, - "summary": { "en": "A day of your AI usage printed as a receipt: its cost, requests and streak, the day against your average day, each agent and its models, and its tokens. Copy it as a picture or save it as a PNG. Reads only your usage by day.", "zh": "把一天的 AI 用量打印成一张小票:费用、请求数、连续使用天数、与日均的倍数、每个 Agent 和它用的模型,以及 token 用量。可复制为图片或保存为 PNG。只读取按天的用量。" } }, - { "package": "@magpie-community/page-fingerprint", "name": "Fingerprint", "kind": "page", "community": true, - "summary": { "en": "Check for suspected model substitution by comparing response fingerprints across Magpie providers. Uses lm-detector with shared challenges, local evidence history and JSON export; sampling uses provider quota. Reference data is downloaded on demand from lm.ikale.io, falling back to cdn.jsdmirror.com.", "zh": "在 Magpie 内检测响应指纹,为疑似模型替换提供核查线索。基于 lm-detector,支持跨渠道共享挑战、保留本地证据与 JSON 导出;采样会消耗模型额度。参考库按需从 lm.ikale.io 下载,备用源为 cdn.jsdmirror.com。" } }, - { "package": "@magpie-community/page-feishu-sidebar", "name": "Feishu sidebar", "kind": "page", "community": true, - "summary": { "en": "Generate an AppLink to open Magpie Web in the Feishu desktop sidebar. The page uses the Web address you enter and reads no Magpie data.", "zh": "生成飞书链接,在飞书桌面端左侧导航打开 Magpie Web。页面使用你填写的 Web 地址,不读取 Magpie 数据。" } }, - { "package": "@magpie-community/page-chat", "name": "Chat", "kind": "page", "community": true, - "summary": { "en": "A chat page in magpie's window by Magic_zero: talk to magpie's own models with your MCP servers, instructions and skills, images and LaTeX. Runs a local service on 127.0.0.1 that only its own page can use.", "zh": "Magic_zero 做的 magpie 窗口内「对话」页面:用 magpie 自己的模型聊天,带上你的 MCP 服务器、指令和技能,支持图片与 LaTeX。本地服务只在 127.0.0.1,只给自己的页面用。" } }, - { "package": "@magpie-community/page-agent-office", "name": "Agent Office", "kind": "page", "community": true, - "summary": { "en": "Reads the session files of Claude Code, Codex, Copilot CLI, Kimi Code, DeepSeek Harness, Gemini CLI, Qwen Code, Pi, omp, Qoder, WorkBuddy, CodeBuddy and Grok Build on this computer (prompts, replies and tool commands; text can be turned off). Shows them in a real WebGL 3D office with eleven scenes, an isometric/2D office, or a grouped workbench with Today usage, recent conversations and approvals.", "zh": "读取本机 Claude Code、Codex、Copilot CLI、Kimi Code、DeepSeek Harness、Gemini CLI、Qwen Code、Pi、omp、Qoder、WorkBuddy、CodeBuddy、Grok Build 的会话文件(含提问、回复与工具命令,可关闭文本),用真 3D 办公室(十一种场景)、等距/2D 办公室或分组工作台展示,带今日用量、最近对话与批准状态。" } }, - { "package": "@magpie-community/page-usage-tree", "name": "Account usage", "kind": "page", "community": true, - "summary": { "en": "A provider -> account -> model usage tree with per-account calls, errors and cost. Adds the account layer that Usage Glance's usage:summary does not contain, read from Magpie's local usage API.", "zh": "按供应商 -> 账号 -> 模型展示用量,包含每个账号的调用、错误和费用;补充 Usage Glance 的 usage:summary 所没有的账号层,数据来自 Magpie 本地用量 API。" } }, - {"package": "@magpie-community/agent-kimi-desktop", "name": "Kimi", "kind": "agent", "community": true, "summary": {"en": "Kimi desktop Work with Magpie models: configure primary and secondary models on its embedded page, select native subagent roles, and disconnect to restore the original service. Not Kimi Code CLI.", "zh": "普通 Kimi 桌面版 Work 使用 Magpie 模型:在内嵌页面配置主、副模型,为原生子代理选择角色,断开时恢复原服务。不是 Kimi Code CLI。"}}, - {"package": "@magpie-community/agent-minimax-design", "name": "MiniMax Design", "kind": "agent", "community": true, "icon": "https://design.minimax.io/favicon.ico", "summary": {"en": "MiniMax Design (the local multimodal creation studio) on magpie's Agents page: one switch points Design's custom text model at magpie's gateway, and Design's model menu follows magpie's model list automatically. Quit Design before switching; Windows only (see COMPATIBILITY.md).", "zh": "把 MiniMax Design(本地多模态创作工作室)加到 magpie 的 Agent 页:一个开关把 Design 的自定义文本模型指到 magpie 网关,Design 的模型菜单随 magpie 的模型列表自动同步。切换前先退出 Design;仅支持 Windows,详见 COMPATIBILITY.md。"}}, - { "package": "@magpie-community/gateway-notify", "name": "Notify", "kind": "middleware", "community": true, - "summary": { "en": "A push when an agent's turn is done or fails, after it ran long enough to be worth one: ntfy, Bark, Gotify, Server酱, Feishu, DingTalk, WeCom, Telegram or your own webhook. Also pushes when a subscription account is signed out and needs signing in again. Agent, project, model and time only by default; the reply's text is opt-in.", "zh": "agent 一轮任务跑完或失败时推送提醒(只提醒跑够设定时长的任务):支持 ntfy、Bark、Gotify、Server酱、飞书、钉钉、企业微信、Telegram 和自定义 webhook。订阅账号掉线需要重新登录时也会推送。默认只发 agent、项目、模型和用时,回复正文需手动开启。" } }, - { "package": "@magpie-community/hooks-mail-watch", "name": "Mail Watch", "kind": "middleware", "community": true, - "summary": { "en": "Email alerts when a provider call fails or a quota window runs out: your own SMTP account, your own words in the mail, and your own list of what to watch, with cooldowns and an hourly cap so a bad hour stays bounded — plus an optional periodic usage digest.", "zh": "供应商调用失败或额度窗口用尽时发邮件提醒:用你自己的 SMTP 账号、自己的邮件文案、自己勾选的监控事项;同类事件冷却合并,另有每小时上限,糟糕的一小时不会失控;另可开启定期用量摘要。" } }, - { "package": "@magpie-community/opencode-cloudflare-workers-ai-auth", "name": "Cloudflare Workers AI", "providers": ["cloudflare-workers-ai"], "community": true, - "summary": { "en": "Cloudflare Workers AI with your API token (Workers AI:Edit): 28+ models discovered live from models.dev, the 8 paid-only models flagged, and a Neurons/day allowance row in magpie quota.", "zh": "用你的 Cloudflare API Token(Workers AI:Edit 权限)登录 Workers AI:28+ 模型经 models.dev 实时发现,付费专属模型自动标注,magpie quota 显示每日 Neurons 免费额度。" } }, - { "package": "@magpie-community/opencode-openai-decisions-auth", "name": "OpenAI Decisions", "providers": ["openai-decisions"], "community": true, "icon": "data:image/svg+xml,%3Csvg%20xmlns%3D%22http%3A%2F%2Fwww.w3.org%2F2000%2Fsvg%22%20viewBox%3D%220%200%2064%2064%22%3E%3Crect%20width%3D%2264%22%20height%3D%2264%22%20rx%3D%2214%22%20fill%3D%22%23111827%22%2F%3E%3Cpath%20d%3D%22M20%2032h12V20h12M32%2032v12h12%22%20fill%3D%22none%22%20stroke%3D%22white%22%20stroke-width%3D%225%22%20stroke-linecap%3D%22round%22%20stroke-linejoin%3D%22round%22%2F%3E%3Ccircle%20cx%3D%2220%22%20cy%3D%2232%22%20r%3D%225%22%20fill%3D%22white%22%2F%3E%3Ccircle%20cx%3D%2246%22%20cy%3D%2220%22%20r%3D%225%22%20fill%3D%22white%22%2F%3E%3Ccircle%20cx%3D%2246%22%20cy%3D%2244%22%20r%3D%225%22%20fill%3D%22white%22%2F%3E%3C%2Fsvg%3E", "summary": { "en": "OpenAI's native Decisions API as System One for Magpie classifiers and decision groups: choice, score and boolean probabilities, with standard API-key sign-in and per-question refusal handling.", "zh": "将 OpenAI 原生 Decisions API 转为 System One,供 Magpie 分类器与决策模型组使用,保留选择、评分和布尔概率;使用标准 API key 登录,单题拒绝时保留其余答案。" } }, - { "package": "@magpie-community/opencode-river-auth", "name": "River", "providers": ["river"], "community": true, - "summary": { "en": "River (river.ai) pay-as-you-go inference with your API key: per-account models from /v1/models, documented reasoning efforts and live prices from River's public pricing endpoint.", "zh": "用你的 API key 接入 River (river.ai) 按量计费推理:按账号从 /v1/models 列出模型,带官方推理档位与实时价格。" } }, - { "package": "@magpie-community/opencode-sap-aicore-auth", "name": "SAP AI Core", "providers": ["sap-aicore"], "community": true, - "summary": { "en": "Existing SAP AI Core deployments with client-secret credential-file sign-in, per-resource-group GPT Responses and Claude Messages discovery, streaming and agent tools. No deployment provisioning or quota reporting; Magpie only.", "zh": "用本机 client-secret 凭据文件连接已有 SAP AI Core 部署:按资源组发现 GPT Responses 与 Claude Messages 模型,支持流式回复和 agent 工具。不创建部署,不报告额度;仅验证 Magpie。" } }, - { "package": "magpie-plugin-pink", "name": "PinK", "kind": "agent", + "summary": { + "en": "An explicit seed copies Magpie's full model catalog into Junie CLI and uses Magpie's selected model as the launch default, without LiteLLM. Native ECO is unavailable for custom profiles in Junie 3893.6.", + "zh": "手动 seed 向 Junie CLI 同步全部模型,并将 Magpie 所选模型设为启动默认模型,无需 LiteLLM。Junie 3893.6 的自定义 profile 不支持原生 ECO。" + } + }, + { + "package": "@magpie-community/middleware-reserve", + "name": "Reserve", + "kind": "middleware", + "community": true, + "summary": { + "en": "Two routings that keep something in reserve: fill accounts up to a ceiling of their allowance, or use subscriptions before API keys.", + "zh": "两种留余量的路由:账号用到额度上限就换下一个,或订阅账号优先于 API key。" + } + }, + { + "package": "@magpie-community/tools-search1api", + "name": "Search1API", + "icon": "https://s1.dev/icon.png", + "kind": "tools", + "community": true, + "summary": { + "en": "Search1API web search, news, agentic search, page retrieval, sitemaps and trends through Magpie's tools MCP server. Sign in with OAuth or an API key.", + "zh": "通过 Magpie 的工具 MCP 服务使用 Search1API 网页搜索、新闻、智能搜索、页面读取、站点链接和趋势;支持 OAuth 或 API key 登录。" + } + }, + { + "package": "@guotree/magpie-websearch", + "name": "magpie-websearch", + "icon": "https://raw.githubusercontent.com/guotree/magpie-websearch/main/assets/icon-128.png", + "kind": "tools", + "community": true, + "summary": { + "en": "Adds 19 search APIs to Magpie: AnySearch, Bocha, Jina, Kagi, Keenable, Linkup, Metaso, Moonshot, Moonshot CN, Ollama, Parallel, Perplexity, Querit, SerpApi, Serper, StepFun, TinyFish, Z.ai, Zhipu. 4 support keyless access; 15 require keys.", + "zh": "为 Magpie 网页搜索增加 19 家搜索源:AnySearch、Bocha、Jina、Kagi、Keenable、Linkup、Metaso、Moonshot、Moonshot CN、Ollama、Parallel、Perplexity、Querit、SerpApi、Serper、StepFun、TinyFish、Z.ai 和智谱。4 家支持免 Key,其余 15 家需要 API Key。" + } + }, + { + "package": "magpie-search-mix", + "name": "magpie-search-mix", + "kind": "tools", + "community": true, + "summary": { + "en": "Mixes two search APIs for the AI: more accurate than one. Try DeepSeek + Tavily or You.com.", + "zh": "混合搜索:同时问两家搜索 API,比只用一家更准。如 DeepSeek 搭配 Tavily 或 You.com。" + } + }, + { + "package": "magpie-comfyui-link", + "name": "ComfyUI", + "kind": "tools", + "community": true, + "summary": { + "en": "A bridge between Magpie agents and locally hosted ComfyUI models, exposing native workflows, model resources and generated images through Magpie's tools MCP endpoint.", + "zh": "连接 Magpie Agent 与本地 ComfyUI 模型的桥接层,通过 Magpie 工具 MCP 入口提供原生工作流、模型资源和生成图片。" + } + }, + { + "package": "@magpie-community/tools-specstream", + "name": "SpecStream", + "kind": "tools", + "community": true, + "summary": { + "en": "SpecStream web search and page extraction for every agent: answers with the source page of each value, and the text of up to 20 URLs a call. Needs a SpecStream API key.", + "zh": "给每个 agent 的 SpecStream 网页搜索和网页提取:每个值都带来源页面,一次最多读取 20 个网址的正文。需要 SpecStream API key。" + } + }, + { + "package": "@magpie-community/tools-calc", + "name": "Calc", + "kind": "tools", + "community": true, + "summary": { + "en": "Exact arithmetic, unit conversion and calendar date math as tools for every agent: no network, no key.", + "zh": "给每个 agent 的精确计算、单位换算和日历日期计算工具:不联网,不需要 key。" + } + }, + { + "package": "@magpie-community/tools-layerfs", + "name": "LayerFS", + "kind": "tools", + "community": true, + "summary": { + "en": "LayerFS workspaces for every agent: deduplicated checkpoints and branches behind stable filesystem paths, instead of a fresh copy of the tree for every agent.", + "zh": "给每个 agent 的 LayerFS 工作区:稳定的文件系统路径背后是可去重的检查点与分支,而不是每个 agent 各复制一棵项目树。" + } + }, + { + "package": "@magpie-community/page-usage", + "name": "Usage Glance", + "kind": "page", + "community": true, + "summary": { + "en": "A page in magpie's window with calls, errors, tokens and cost for today, 7 or 30 days, and the models behind them. Reads only your usage summary.", + "zh": "在 magpie 窗口里一眼看到今天、7 天或 30 天的调用、错误、token、费用和对应模型。只读取用量汇总。" + } + }, + { + "package": "@magpie-community/page-amp-usage", + "name": "Amp Usage", + "kind": "page", + "community": true, + "icon": "https://raw.githubusercontent.com/magpie-community/plugins/main/packages/amp-usage/page/icon.svg", + "summary": { + "en": "An Amp usage board: requests, tokens and cache hit rate by hour, day and model, read from your Amp account and legacy local threads. Its background service syncs that usage into Magpie requests and Day card, with duplicate checks; set sync to false to only view it.", + "zh": "Amp 用量看板:请求数、Token、缓存命中率,按小时、按日和按模型查看,数据来自 Amp 账号和旧版本地线程。后台服务把这些用量同步进 Magpie 请求统计和每日卡片,带去重检查;把 sync 设为 false 则只查看。" + } + }, + { + "package": "@magpie-community/page-day-card", + "name": "Day Card", + "kind": "page", + "community": true, + "summary": { + "en": "A day of your AI usage printed as a receipt: its cost, requests and streak, the day against your average day, each agent and its models, and its tokens. Copy it as a picture or save it as a PNG. Reads only your usage by day.", + "zh": "把一天的 AI 用量打印成一张小票:费用、请求数、连续使用天数、与日均的倍数、每个 Agent 和它用的模型,以及 token 用量。可复制为图片或保存为 PNG。只读取按天的用量。" + } + }, + { + "package": "@magpie-community/page-fingerprint", + "name": "Fingerprint", + "kind": "page", + "community": true, + "summary": { + "en": "Check for suspected model substitution by comparing response fingerprints across Magpie providers. Uses lm-detector with shared challenges, local evidence history and JSON export; sampling uses provider quota. Reference data is downloaded on demand from lm.ikale.io, falling back to cdn.jsdmirror.com.", + "zh": "在 Magpie 内检测响应指纹,为疑似模型替换提供核查线索。基于 lm-detector,支持跨渠道共享挑战、保留本地证据与 JSON 导出;采样会消耗模型额度。参考库按需从 lm.ikale.io 下载,备用源为 cdn.jsdmirror.com。" + } + }, + { + "package": "@magpie-community/page-feishu-sidebar", + "name": "Feishu sidebar", + "kind": "page", + "community": true, + "summary": { + "en": "Generate an AppLink to open Magpie Web in the Feishu desktop sidebar. The page uses the Web address you enter and reads no Magpie data.", + "zh": "生成飞书链接,在飞书桌面端左侧导航打开 Magpie Web。页面使用你填写的 Web 地址,不读取 Magpie 数据。" + } + }, + { + "package": "@magpie-community/page-chat", + "name": "Chat", + "kind": "page", + "community": true, + "summary": { + "en": "A chat page in magpie's window by Magic_zero: talk to magpie's own models with your MCP servers, instructions and skills, images and LaTeX. Runs a local service on 127.0.0.1 that only its own page can use.", + "zh": "Magic_zero 做的 magpie 窗口内「对话」页面:用 magpie 自己的模型聊天,带上你的 MCP 服务器、指令和技能,支持图片与 LaTeX。本地服务只在 127.0.0.1,只给自己的页面用。" + } + }, + { + "package": "@magpie-community/page-agent-office", + "name": "Agent Office", + "kind": "page", + "community": true, + "summary": { + "en": "Reads the session files of Claude Code, Codex, Copilot CLI, Kimi Code, DeepSeek Harness, Gemini CLI, Qwen Code, Pi, omp, Qoder, WorkBuddy, CodeBuddy and Grok Build on this computer (prompts, replies and tool commands; text can be turned off). Shows them in a real WebGL 3D office with eleven scenes, an isometric/2D office, or a grouped workbench with Today usage, recent conversations and approvals.", + "zh": "读取本机 Claude Code、Codex、Copilot CLI、Kimi Code、DeepSeek Harness、Gemini CLI、Qwen Code、Pi、omp、Qoder、WorkBuddy、CodeBuddy、Grok Build 的会话文件(含提问、回复与工具命令,可关闭文本),用真 3D 办公室(十一种场景)、等距/2D 办公室或分组工作台展示,带今日用量、最近对话与批准状态。" + } + }, + { + "package": "@magpie-community/page-usage-tree", + "name": "Account usage", + "kind": "page", + "community": true, + "summary": { + "en": "A provider -> account -> model usage tree with per-account calls, errors and cost. Adds the account layer that Usage Glance's usage:summary does not contain, read from Magpie's local usage API.", + "zh": "按供应商 -> 账号 -> 模型展示用量,包含每个账号的调用、错误和费用;补充 Usage Glance 的 usage:summary 所没有的账号层,数据来自 Magpie 本地用量 API。" + } + }, + { + "package": "@magpie-community/agent-kimi-desktop", + "name": "Kimi", + "kind": "agent", + "community": true, + "summary": { + "en": "Kimi desktop Work with Magpie models: configure primary and secondary models on its embedded page, select native subagent roles, and disconnect to restore the original service. Not Kimi Code CLI.", + "zh": "普通 Kimi 桌面版 Work 使用 Magpie 模型:在内嵌页面配置主、副模型,为原生子代理选择角色,断开时恢复原服务。不是 Kimi Code CLI。" + } + }, + { + "package": "@magpie-community/agent-minimax-design", + "name": "MiniMax Design", + "kind": "agent", + "community": true, + "icon": "https://design.minimax.io/favicon.ico", + "summary": { + "en": "MiniMax Design (the local multimodal creation studio) on magpie's Agents page: one switch points Design's custom text model at magpie's gateway, and Design's model menu follows magpie's model list automatically. Quit Design before switching; Windows only (see COMPATIBILITY.md).", + "zh": "把 MiniMax Design(本地多模态创作工作室)加到 magpie 的 Agent 页:一个开关把 Design 的自定义文本模型指到 magpie 网关,Design 的模型菜单随 magpie 的模型列表自动同步。切换前先退出 Design;仅支持 Windows,详见 COMPATIBILITY.md。" + } + }, + { + "package": "@magpie-community/gateway-notify", + "name": "Notify", + "kind": "middleware", + "community": true, + "summary": { + "en": "A push when an agent's turn is done or fails, after it ran long enough to be worth one: ntfy, Bark, Gotify, Server酱, Feishu, DingTalk, WeCom, Telegram or your own webhook. Also pushes when a subscription account is signed out and needs signing in again. Agent, project, model and time only by default; the reply's text is opt-in.", + "zh": "agent 一轮任务跑完或失败时推送提醒(只提醒跑够设定时长的任务):支持 ntfy、Bark、Gotify、Server酱、飞书、钉钉、企业微信、Telegram 和自定义 webhook。订阅账号掉线需要重新登录时也会推送。默认只发 agent、项目、模型和用时,回复正文需手动开启。" + } + }, + { + "package": "@magpie-community/hooks-mail-watch", + "name": "Mail Watch", + "kind": "middleware", + "community": true, + "summary": { + "en": "Email alerts when a provider call fails or a quota window runs out: your own SMTP account, your own words in the mail, and your own list of what to watch, with cooldowns and an hourly cap so a bad hour stays bounded — plus an optional periodic usage digest.", + "zh": "供应商调用失败或额度窗口用尽时发邮件提醒:用你自己的 SMTP 账号、自己的邮件文案、自己勾选的监控事项;同类事件冷却合并,另有每小时上限,糟糕的一小时不会失控;另可开启定期用量摘要。" + } + }, + { + "package": "@magpie-community/opencode-cloudflare-workers-ai-auth", + "name": "Cloudflare Workers AI", + "providers": [ + "cloudflare-workers-ai" + ], + "community": true, + "summary": { + "en": "Cloudflare Workers AI with your API token (Workers AI:Edit): 28+ models discovered live from models.dev, the 8 paid-only models flagged, and a Neurons/day allowance row in magpie quota.", + "zh": "用你的 Cloudflare API Token(Workers AI:Edit 权限)登录 Workers AI:28+ 模型经 models.dev 实时发现,付费专属模型自动标注,magpie quota 显示每日 Neurons 免费额度。" + } + }, + { + "package": "@magpie-community/opencode-openai-decisions-auth", + "name": "OpenAI Decisions", + "providers": [ + "openai-decisions" + ], + "community": true, + "icon": "data:image/svg+xml,%3Csvg%20xmlns%3D%22http%3A%2F%2Fwww.w3.org%2F2000%2Fsvg%22%20viewBox%3D%220%200%2064%2064%22%3E%3Crect%20width%3D%2264%22%20height%3D%2264%22%20rx%3D%2214%22%20fill%3D%22%23111827%22%2F%3E%3Cpath%20d%3D%22M20%2032h12V20h12M32%2032v12h12%22%20fill%3D%22none%22%20stroke%3D%22white%22%20stroke-width%3D%225%22%20stroke-linecap%3D%22round%22%20stroke-linejoin%3D%22round%22%2F%3E%3Ccircle%20cx%3D%2220%22%20cy%3D%2232%22%20r%3D%225%22%20fill%3D%22white%22%2F%3E%3Ccircle%20cx%3D%2246%22%20cy%3D%2220%22%20r%3D%225%22%20fill%3D%22white%22%2F%3E%3Ccircle%20cx%3D%2246%22%20cy%3D%2244%22%20r%3D%225%22%20fill%3D%22white%22%2F%3E%3C%2Fsvg%3E", + "summary": { + "en": "OpenAI's native Decisions API as System One for Magpie classifiers and decision groups: choice, score and boolean probabilities, with standard API-key sign-in and per-question refusal handling.", + "zh": "将 OpenAI 原生 Decisions API 转为 System One,供 Magpie 分类器与决策模型组使用,保留选择、评分和布尔概率;使用标准 API key 登录,单题拒绝时保留其余答案。" + } + }, + { + "package": "@magpie-community/opencode-openrouter-decisions-auth", + "name": "OpenRouter Decisions", + "providers": [ + "openrouter-decisions" + ], + "community": true, + "icon": "data:image/svg+xml,%3Csvg%20xmlns%3D%22http%3A%2F%2Fwww.w3.org%2F2000%2Fsvg%22%20viewBox%3D%220%200%2064%2064%22%3E%3Ccircle%20cx%3D%2232%22%20cy%3D%2232%22%20r%3D%2229%22%20fill%3D%22%230f172a%22%2F%3E%3Cpath%20d%3D%22M20%2040a12%2012%200%201%201%2012%2012%22%20fill%3D%22none%22%20stroke%3D%22%2365a3ff%22%20stroke-width%3D%225%22%20stroke-linecap%3D%22round%22%2F%3E%3Ccircle%20cx%3D%2244%22%20cy%3D%2228%22%20r%3D%226%22%20fill%3D%22%2365a3ff%22%2F%3E%3C%2Fsvg%3E", + "summary": { + "en": "OpenRouter's Decisions API as System One for Magpie classifiers and decision groups: Microsoft Decision 1.1 and 1, with choice, score and boolean probabilities, credit usage and per-question refusal handling.", + "zh": "将 OpenRouter Decisions API 转为 System One,供 Magpie 分类器与决策模型组使用:Microsoft Decision 1.1 与 1,保留选择、评分和布尔概率、额度用量,单题拒绝时保留其余答案。" + } + }, + { + "package": "@magpie-community/opencode-river-auth", + "name": "River", + "providers": [ + "river" + ], + "community": true, + "summary": { + "en": "River (river.ai) pay-as-you-go inference with your API key: per-account models from /v1/models, documented reasoning efforts and live prices from River's public pricing endpoint.", + "zh": "用你的 API key 接入 River (river.ai) 按量计费推理:按账号从 /v1/models 列出模型,带官方推理档位与实时价格。" + } + }, + { + "package": "@magpie-community/opencode-sap-aicore-auth", + "name": "SAP AI Core", + "providers": [ + "sap-aicore" + ], + "community": true, + "summary": { + "en": "Existing SAP AI Core deployments with client-secret credential-file sign-in, per-resource-group GPT Responses and Claude Messages discovery, streaming and agent tools. No deployment provisioning or quota reporting; Magpie only.", + "zh": "用本机 client-secret 凭据文件连接已有 SAP AI Core 部署:按资源组发现 GPT Responses 与 Claude Messages 模型,支持流式回复和 agent 工具。不创建部署,不报告额度;仅验证 Magpie。" + } + }, + { + "package": "magpie-plugin-pink", + "name": "PinK", + "kind": "agent", "icon": "https://raw.githubusercontent.com/YuchenWell/magpie-plugin-pink/main/assets/pink.png", - "summary": { "en": "Connect PinK to Magpie's models from the Agents page and track gateway usage and cost. Disconnecting restores the managed configuration.", "zh": "在 Agent 页面连接 PinK,使用 Magpie 中配置的模型并查看调用用量与费用;断开时恢复插件管理的配置。" } }, - { "package": "@hoptoper/magpie-agent-wispterm", "name": "WispTerm", "kind": "agent", - "summary": { "en": "WispTerm on magpie's Agents page: pick one of magpie's models for its Agent, switch reasoning effort with /effort, and switch it off to put its config back. Needs WispTerm 1.46.1 or later.", "zh": "把 WispTerm 加到 magpie 的 Agent 页:给它的 Agent 选 magpie 的模型,用 /effort 切换思考强度,关掉即恢复原来的配置。需要 WispTerm 1.46.1 或更新版本。" } }, - { "package": "@hoptoper/magpie-agent-citebox", "name": "CiteBox", "kind": "agent", - "summary": { "en": "CiteBox on magpie's Agents page: pick one of magpie's models for its text AI, set reasoning effort in Settings, and switch it off to put its config back. Needs CiteBox 0.38.0 or later.", "zh": "把 CiteBox 加到 magpie 的 Agent 页:给它的文本 AI 选 magpie 的模型,在设置里调整思考强度,关掉即恢复原来的配置。需要 CiteBox 0.38.0 或更新版本。" } }, - { "package": "@hoptoper/magpie-agent-wisp-science", "name": "Wisp Science", "kind": "agent", - "summary": { "en": "Wisp Science on magpie's Agents page: pick one of magpie's models, set reasoning effort in the composer, and switch it off to put its config back. Needs Wisp Science 1.23.0 or later.", "zh": "把 Wisp Science 加到 magpie 的 Agent 页:选择 magpie 的模型,在输入框里设置思考强度,关掉即恢复原来的配置。需要 Wisp Science 1.23.0 或更新版本。" } }, - { "package": "magpie-agent-kaku", "name": "Kaku", "kind": "agent", + "summary": { + "en": "Connect PinK to Magpie's models from the Agents page and track gateway usage and cost. Disconnecting restores the managed configuration.", + "zh": "在 Agent 页面连接 PinK,使用 Magpie 中配置的模型并查看调用用量与费用;断开时恢复插件管理的配置。" + } + }, + { + "package": "@hoptoper/magpie-agent-wispterm", + "name": "WispTerm", + "kind": "agent", + "summary": { + "en": "WispTerm on magpie's Agents page: pick one of magpie's models for its Agent, switch reasoning effort with /effort, and switch it off to put its config back. Needs WispTerm 1.46.1 or later.", + "zh": "把 WispTerm 加到 magpie 的 Agent 页:给它的 Agent 选 magpie 的模型,用 /effort 切换思考强度,关掉即恢复原来的配置。需要 WispTerm 1.46.1 或更新版本。" + } + }, + { + "package": "@hoptoper/magpie-agent-citebox", + "name": "CiteBox", + "kind": "agent", + "summary": { + "en": "CiteBox on magpie's Agents page: pick one of magpie's models for its text AI, set reasoning effort in Settings, and switch it off to put its config back. Needs CiteBox 0.38.0 or later.", + "zh": "把 CiteBox 加到 magpie 的 Agent 页:给它的文本 AI 选 magpie 的模型,在设置里调整思考强度,关掉即恢复原来的配置。需要 CiteBox 0.38.0 或更新版本。" + } + }, + { + "package": "@hoptoper/magpie-agent-wisp-science", + "name": "Wisp Science", + "kind": "agent", + "summary": { + "en": "Wisp Science on magpie's Agents page: pick one of magpie's models, set reasoning effort in the composer, and switch it off to put its config back. Needs Wisp Science 1.23.0 or later.", + "zh": "把 Wisp Science 加到 magpie 的 Agent 页:选择 magpie 的模型,在输入框里设置思考强度,关掉即恢复原来的配置。需要 Wisp Science 1.23.0 或更新版本。" + } + }, + { + "package": "magpie-agent-kaku", + "name": "Kaku", + "kind": "agent", "icon": "https://raw.githubusercontent.com/shlroland/magpie-agent-kaku/main/assets/icon.png", - "summary": { "en": "Kaku (the terminal) Assistant on magpie's Agents page: one switch points it at magpie's gateway, with a routing group for each of its roles: Simple for # commands and command fixes, Deep for the Cmd+L chat and tools, and optional Memory. Switching it off restores assistant.toml byte for byte. macOS.", "zh": "把 Kaku 终端的 AI 助手加到 magpie 的 Agent 页:一个开关把它接到 magpie 网关,并给每个角色绑定一个路由分组:Simple 负责 # 命令和报错修复,Deep 负责 Cmd+L 对话和工具调用,Memory 可选。关掉即逐字节恢复 assistant.toml。仅 macOS。" } }, - { "package": "@magpie-community/agent-eta", "name": "Eta", "kind": "agent", "community": true, "icon": "data:image/svg+xml;base64,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", - "summary": { "en": "Use Magpie models and routing groups in Eta’s own model picker. Requires Eta 0.1.164 or later; restart Eta after connecting.", "zh": "在 Eta 自带的模型选择器中使用 Magpie 模型与路由组。需要 Eta 0.1.164 或更新版本;连接后重启 Eta。" } }, - { "package": "@magpie-community/agent-rhino9-ai", "name": "Magpie-Rhino", "kind": "agent", "community": true, - "summary": { "en": "Magpie-Rhino brings Magpie models, custom groups and strategies into Rhino 9's AI panel through Rhino MCP. Install the companion .yak, restart Rhino, then open its AI panel and choose Magpie; no command required. Local gateway only.", "zh": "Magpie-Rhino 将模型、自定义组和策略接入 Rhino 9 原生 AI 面板,通过 Rhino MCP 建模。安装配套 .yak 并重启后,直接打开 AI 面板选择 Magpie,无需输入命令。仅连接本机网关。" } }, - { "package": "@magpie-community/tools-goanyapi", "name": "GoAnyAPI", "kind": "tools", "community": true, - "summary": { "en": "Explore Google Trends, website traffic, domain rating (DR), keyword suggestions, SERPs, backlinks, and ads statistics for market, keyword, and competitor research.", "zh": "查询 Google 趋势、网站流量、域名权重(DR)、关键词建议、SERP、外链和广告统计,辅助市场、关键词与竞品研究。" } }, + "summary": { + "en": "Kaku (the terminal) Assistant on magpie's Agents page: one switch points it at magpie's gateway, with a routing group for each of its roles: Simple for # commands and command fixes, Deep for the Cmd+L chat and tools, and optional Memory. Switching it off restores assistant.toml byte for byte. macOS.", + "zh": "把 Kaku 终端的 AI 助手加到 magpie 的 Agent 页:一个开关把它接到 magpie 网关,并给每个角色绑定一个路由分组:Simple 负责 # 命令和报错修复,Deep 负责 Cmd+L 对话和工具调用,Memory 可选。关掉即逐字节恢复 assistant.toml。仅 macOS。" + } + }, + { + "package": "@magpie-community/agent-eta", + "name": "Eta", + "kind": "agent", + "community": true, + "icon": "data:image/svg+xml;base64,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", + "summary": { + "en": "Use Magpie models and routing groups in Eta’s own model picker. Requires Eta 0.1.164 or later; restart Eta after connecting.", + "zh": "在 Eta 自带的模型选择器中使用 Magpie 模型与路由组。需要 Eta 0.1.164 或更新版本;连接后重启 Eta。" + } + }, + { + "package": "@magpie-community/agent-rhino9-ai", + "name": "Magpie-Rhino", + "kind": "agent", + "community": true, + "summary": { + "en": "Magpie-Rhino brings Magpie models, custom groups and strategies into Rhino 9's AI panel through Rhino MCP. Install the companion .yak, restart Rhino, then open its AI panel and choose Magpie; no command required. Local gateway only.", + "zh": "Magpie-Rhino 将模型、自定义组和策略接入 Rhino 9 原生 AI 面板,通过 Rhino MCP 建模。安装配套 .yak 并重启后,直接打开 AI 面板选择 Magpie,无需输入命令。仅连接本机网关。" + } + }, + { + "package": "@magpie-community/tools-goanyapi", + "name": "GoAnyAPI", + "kind": "tools", + "community": true, + "summary": { + "en": "Explore Google Trends, website traffic, domain rating (DR), keyword suggestions, SERPs, backlinks, and ads statistics for market, keyword, and competitor research.", + "zh": "查询 Google 趋势、网站流量、域名权重(DR)、关键词建议、SERP、外链和广告统计,辅助市场、关键词与竞品研究。" + } + }, { "package": "@magpie-community/opencode-orcaterm-auth", "name": "OrcaTerm", @@ -211,8 +1138,17 @@ "zh": "用腾讯云遨驰终端(OrcaTerm)AI 助手的模型——DeepSeek、混元、GLM、Kimi——走账号每月问答额度。非官方,条款风险见 README。" } }, - { "package": "magpie-aws-bedrock-sso", "name": "AWS Bedrock SSO", "providers": ["bedrock-sso"], - "summary": { "en": "Use Amazon Bedrock Claude and GPT models with your AWS CLI profile, including SSO / IAM Identity Center. SigV4 authentication; no Bedrock API key needed.", "zh": "使用 AWS CLI 配置(支持 SSO / IAM Identity Center)接入 Amazon Bedrock 的 Claude 和 GPT 模型,通过 SigV4 签名,无需 Bedrock API 密钥。" } }, + { + "package": "magpie-aws-bedrock-sso", + "name": "AWS Bedrock SSO", + "providers": [ + "bedrock-sso" + ], + "summary": { + "en": "Use Amazon Bedrock Claude and GPT models with your AWS CLI profile, including SSO / IAM Identity Center. SigV4 authentication; no Bedrock API key needed.", + "zh": "使用 AWS CLI 配置(支持 SSO / IAM Identity Center)接入 Amazon Bedrock 的 Claude 和 GPT 模型,通过 SigV4 签名,无需 Bedrock API 密钥。" + } + }, { "package": "magpie-providers-hub", "name": "Providers Hub", @@ -256,19 +1192,90 @@ "zh": "窗口预热(原「定时消息」):提前开启订阅的 5 小时或每周额度窗口,给额度保活。给指定的供应商、账号和模型发一条固定短消息(\"hi\",max_tokens 16),可在窗口重置后、每天定时或每隔几小时发送。每次发送都会消耗该供应商的额度(配额)。在本机 127.0.0.1 跑一个本地服务,只调用 magpie 自己的网关;令牌只交给自己的页面。" } }, - { "package": "@magpie-community/opencode-astudio-auth", "name": "AStudio", "providers": ["astudio"], "community": true, - "summary": { "en": "Unofficial AStudio (iFLYTEK Astron Studio): sign in through a browser (each account gets its own pool slot), reuse the desktop sign-in, or paste a token, to serve its Responses model gateway with the live model catalog, daily login-points check-in (read from the credit ledger) and points balance with spark-point expiry. Differs from Providers Hub's hub-astudio, which signs in on a local page (SMS, password or WeChat QR) and calls chat/completions; on one account both share the same daily points pool.", "zh": "非官方 AStudio(科大讯飞 Astron Studio):支持浏览器登录(每个账号占账号池一格)、复用桌面端登录态或粘贴凭据,接入其模型网关,含实时模型目录、每日登录积分签到(按积分流水判定)与积分余额/星火到期时间。与 Providers Hub 的 hub-astudio 不同:那边在本地登录页用短信、密码或微信扫码登录,走 chat/completions;同一账号两边共用一个每日积分池。" } }, - { "package": "@magpie-community/page-free-tokens", "name": "Token 羊毛", "kind": "page", "community": true, - "summary": { "en": "AIHOT's Token freebies in magpie's window, in three tiers: already configured and usable, needs one more plugin, or no plugin yet.", "zh": "AIHOT 的 Token 羊毛,在 magpie 窗口里分三档:已配置可用 / 差一个插件 / 暂无插件。" } }, - { "package": "@magpie-community/opencode-agnes-code-auth", "name": "Agnes Code", "providers": ["agnes-code"], "community": true, - "summary": { "en": "Unofficial Agnes Code desktop: reuse the desktop sign-in for its OpenAI-compatible gateway, with the live model catalog, daily credit check-in and points balance. Desktop decryption is Windows-only; a paste-token method covers the rest. Differs from opencode-agnes-auth (`agnes`), which signs in to Agnes AI international in Chrome and meters that product's plans and USD balance: this one serves the Agnes Code desktop product and its daily credits.", "zh": "非官方 Agnes Code 桌面端:复用桌面端登录态接入其 OpenAI 兼容网关,含实时模型目录、每日积分签到与余额显示。桌面端解密仅限 Windows,其余平台用粘贴 accessToken。与 opencode-agnes-auth(`agnes`)不同:那边在 Chrome 里登录 Agnes AI 国际版、显示该产品的套餐与美元余额;本包服务 Agnes Code 桌面端产品与它的每日积分。" } }, - { "package": "@magpie-community/opencode-tabbit-auth", "name": "Tabbit", "icon": "https://web.tabbit.com/favicon.ico", "providers": ["tabbit"], "community": true, - "summary": { "en": "Unofficial Tabbit (web.tabbit.com, and web.tabbit.ai for international accounts) provider: reuse the Tabbit Browser sign-in on this computer to serve its chat models, with the account's live catalog, cycle usage and the daily check-in. No browser automation: a turn is Tabbit's own room → run → event-stream calls. Client tool calls are not available (Tabbit's tools are its own).", "zh": "非官方 Tabbit(web.tabbit.com,国际版账号走 web.tabbit.ai)插件:复用本机 Tabbit 浏览器的登录态接入它的聊天模型,含账号实时模型目录、周期用量与每日签到。不走浏览器自动化:一轮对话就是 Tabbit 自己的 建房间 → 发起 → 事件流。不支持客户端工具调用(Tabbit 的工具是它自己的)。" } }, - { "package": "magpie-plugin-codar", "name": "Codar (科搭)", "icon": "https://codar.speedai.com/assets/codar-icon.png", "providers": ["codar"], "community": true, - "summary": { "en": "Codar (科搭), SpeedAI's research desktop app: its 17 models (the auto router plus the DeepSeek, GLM, GPT and Claude families) over OpenAI Chat and Anthropic Messages, with the model list refreshed live from the vendor. Signs in by reading the desktop app's token file on this machine (read-only) or by pasting an access token; the daily check-in reports the claim and carries the next claim time; the card shows general and agent points and any running free-model activity.", "zh": "科搭(Codar,SpeedAI 出品的研究桌面应用):17 个模型(auto 路由 + DeepSeek、GLM、GPT、Claude 各系),走 OpenAI Chat 与 Anthropic Messages,模型列表从厂商实时刷新。登录可读取本机桌面端令牌(只读)或粘贴访问令牌;每日签到报告领取结果并带上下次可领时间;卡片显示通用与 Agent 点数及进行中的免费模型活动。" } }, - {"package": "@magpie-community/gateway-agentcompass", "name": "AgentCompass", "kind": "middleware", "community": true, "summary": {"en": "AgentCompass routing for Magpie: picks : from candidateModels by a task profile's four capability categories, per-effort time and price, and session continuity, orders pool seats by subscription quota, and optionally serves /v1/systemone for Routing Groups.", "zh": "AgentCompass 路由插件:按任务画像的四类能力需求、各档位的耗时与价格和会话连续性从 candidateModels 选出 :,按订阅配额排号池席位,并可选作为路由组 /v1/systemone 分类器。"}, "icon": "data:image/svg+xml;base64,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"}, - { "package": "@magpie-community/opencode-mclaw-auth", "name": "MClaw", "icon": "https://yun.139.com/w/static/img/LOGO.png", "providers": ["mclaw"], "community": true, - "summary": { "en": "Unofficial MClaw provider: China Mobile Cloud's MClaw agent (ai.yun.139.com) with an OpenAI-compatible endpoint — the browser sign-in harvests the ticket by itself, auto-fetches the llm key and the token multipliers, serves the live model catalogue and the aitokens pool usage, and renews the ticket before it expires.", "zh": "非官方 MClaw 插件:中国移动云 MClaw 智能体(ai.yun.139.com),OpenAI 兼容接口——浏览器登录一次自动收票、自动取 llm key 与积分倍率,动态模型目录与 aitokens 池用量,凭证到期前自动续期。" } }, + { + "package": "@magpie-community/opencode-astudio-auth", + "name": "AStudio", + "providers": [ + "astudio" + ], + "community": true, + "summary": { + "en": "Unofficial AStudio (iFLYTEK Astron Studio): sign in through a browser (each account gets its own pool slot), reuse the desktop sign-in, or paste a token, to serve its Responses model gateway with the live model catalog, daily login-points check-in (read from the credit ledger) and points balance with spark-point expiry. Differs from Providers Hub's hub-astudio, which signs in on a local page (SMS, password or WeChat QR) and calls chat/completions; on one account both share the same daily points pool.", + "zh": "非官方 AStudio(科大讯飞 Astron Studio):支持浏览器登录(每个账号占账号池一格)、复用桌面端登录态或粘贴凭据,接入其模型网关,含实时模型目录、每日登录积分签到(按积分流水判定)与积分余额/星火到期时间。与 Providers Hub 的 hub-astudio 不同:那边在本地登录页用短信、密码或微信扫码登录,走 chat/completions;同一账号两边共用一个每日积分池。" + } + }, + { + "package": "@magpie-community/page-free-tokens", + "name": "Token 羊毛", + "kind": "page", + "community": true, + "summary": { + "en": "AIHOT's Token freebies in magpie's window, in three tiers: already configured and usable, needs one more plugin, or no plugin yet.", + "zh": "AIHOT 的 Token 羊毛,在 magpie 窗口里分三档:已配置可用 / 差一个插件 / 暂无插件。" + } + }, + { + "package": "@magpie-community/opencode-agnes-code-auth", + "name": "Agnes Code", + "providers": [ + "agnes-code" + ], + "community": true, + "summary": { + "en": "Unofficial Agnes Code desktop: reuse the desktop sign-in for its OpenAI-compatible gateway, with the live model catalog, daily credit check-in and points balance. Desktop decryption is Windows-only; a paste-token method covers the rest. Differs from opencode-agnes-auth (`agnes`), which signs in to Agnes AI international in Chrome and meters that product's plans and USD balance: this one serves the Agnes Code desktop product and its daily credits.", + "zh": "非官方 Agnes Code 桌面端:复用桌面端登录态接入其 OpenAI 兼容网关,含实时模型目录、每日积分签到与余额显示。桌面端解密仅限 Windows,其余平台用粘贴 accessToken。与 opencode-agnes-auth(`agnes`)不同:那边在 Chrome 里登录 Agnes AI 国际版、显示该产品的套餐与美元余额;本包服务 Agnes Code 桌面端产品与它的每日积分。" + } + }, + { + "package": "@magpie-community/opencode-tabbit-auth", + "name": "Tabbit", + "icon": "https://web.tabbit.com/favicon.ico", + "providers": [ + "tabbit" + ], + "community": true, + "summary": { + "en": "Unofficial Tabbit (web.tabbit.com, and web.tabbit.ai for international accounts) provider: reuse the Tabbit Browser sign-in on this computer to serve its chat models, with the account's live catalog, cycle usage and the daily check-in. No browser automation: a turn is Tabbit's own room → run → event-stream calls. Client tool calls are not available (Tabbit's tools are its own).", + "zh": "非官方 Tabbit(web.tabbit.com,国际版账号走 web.tabbit.ai)插件:复用本机 Tabbit 浏览器的登录态接入它的聊天模型,含账号实时模型目录、周期用量与每日签到。不走浏览器自动化:一轮对话就是 Tabbit 自己的 建房间 → 发起 → 事件流。不支持客户端工具调用(Tabbit 的工具是它自己的)。" + } + }, + { + "package": "magpie-plugin-codar", + "name": "Codar (科搭)", + "icon": "https://codar.speedai.com/assets/codar-icon.png", + "providers": [ + "codar" + ], + "community": true, + "summary": { + "en": "Codar (科搭), SpeedAI's research desktop app: its 17 models (the auto router plus the DeepSeek, GLM, GPT and Claude families) over OpenAI Chat and Anthropic Messages, with the model list refreshed live from the vendor. Signs in by reading the desktop app's token file on this machine (read-only) or by pasting an access token; the daily check-in reports the claim and carries the next claim time; the card shows general and agent points and any running free-model activity.", + "zh": "科搭(Codar,SpeedAI 出品的研究桌面应用):17 个模型(auto 路由 + DeepSeek、GLM、GPT、Claude 各系),走 OpenAI Chat 与 Anthropic Messages,模型列表从厂商实时刷新。登录可读取本机桌面端令牌(只读)或粘贴访问令牌;每日签到报告领取结果并带上下次可领时间;卡片显示通用与 Agent 点数及进行中的免费模型活动。" + } + }, + { + "package": "@magpie-community/gateway-agentcompass", + "name": "AgentCompass", + "kind": "middleware", + "community": true, + "summary": { + "en": "AgentCompass routing for Magpie: picks : from candidateModels by a task profile's four capability categories, per-effort time and price, and session continuity, orders pool seats by subscription quota, and optionally serves /v1/systemone for Routing Groups.", + "zh": "AgentCompass 路由插件:按任务画像的四类能力需求、各档位的耗时与价格和会话连续性从 candidateModels 选出 :,按订阅配额排号池席位,并可选作为路由组 /v1/systemone 分类器。" + }, + "icon": "data:image/svg+xml;base64,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" + }, + { + "package": "@magpie-community/opencode-mclaw-auth", + "name": "MClaw", + "icon": "https://yun.139.com/w/static/img/LOGO.png", + "providers": [ + "mclaw" + ], + "community": true, + "summary": { + "en": "Unofficial MClaw provider: China Mobile Cloud's MClaw agent (ai.yun.139.com) with an OpenAI-compatible endpoint — the browser sign-in harvests the ticket by itself, auto-fetches the llm key and the token multipliers, serves the live model catalogue and the aitokens pool usage, and renews the ticket before it expires.", + "zh": "非官方 MClaw 插件:中国移动云 MClaw 智能体(ai.yun.139.com),OpenAI 兼容接口——浏览器登录一次自动收票、自动取 llm key 与积分倍率,动态模型目录与 aitokens 池用量,凭证到期前自动续期。" + } + }, { "package": "@magpie-community/middleware-privacy-gate", "name": "ShyPie Privacy Gate", @@ -280,30 +1287,160 @@ "zh": "由 ShyPie 项目维护。ShyPie 面向 DeepSeek Harness 的隐私优先工作流中 Magpie 一侧的出口门:当 DSH 通过 Magpie 网关使用 DeepSeek 模型时,对 Magpie 无法本地掩码的图片/文件/音频、编码数据块和允许列表之外的模型一律拒绝。它不做文本掩码(文本由 Magpie Privacy 负责);图片和文件默认返回 403,除非 localModels 和 localProviders 都允许。设置 binaryPolicy: guide 可暂扣二进制内容并提示模型选择本地提取工具;结构性检查按失败关闭处理;超过 maxNodes 个值的请求会被拒绝,但超过 scanChars 的文本不会被检查(优先扫描最新内容,每个字段有上限),较长的编码串可能被误判。" } }, - { "package": "magpie-deepaicode", "name": "DeepAI Code", "icon": "https://raw.githubusercontent.com/aqiu817/magpie-deepaicode/main/icon.svg", "providers": ["deepaicode"], - "summary": { "en": "Unofficial DeepAI Code access for magpie: browser sign-in, the desktop app's saved credentials or an API key, direct upstream requests with the device's real machine fingerprint, live models, daily Theos quota and check-in.", "zh": "非官方接入 DeepAI Code:浏览器登录、桌面端凭据或 API Key,直连上游并使用真实设备指纹,动态模型目录、每日 Theos 额度与签到。" } }, - { "package": "@magpie-community/opencode-agy-cli-auth", "name": "Antigravity (local agy CLI)", "icon": "antigravity-color", "providers": ["agy-cli"], "community": true, - "summary": { "en": "Unofficial provider that runs the agy CLI you installed and signed in to through the Agent Client Protocol (agy acp), serving only the models agy offers this client. Unlike Antigravity ACP, which downloads Google's ACP server and rides the Antigravity IDE's sign-in, it uses your own agy and adds a process pool, a history budget, a bridge for the client's tools, session resume and agy's usage windows.", "zh": "非官方插件:通过 Agent Client Protocol(agy acp)驱动你自己安装并登录的 agy CLI,只提供 agy 向本客户端开放的模型。与自行下载 Google ACP 服务端、依赖 Antigravity IDE 登录的 Antigravity ACP 不同,它用你本机的 agy,并带进程池、历史体积预算、客户端工具桥接、会话续接与 agy 用量窗口。" } }, - { "package": "opencode-teleagent-auth", "name": "TeleAgent 模型", "icon": "https://raw.githubusercontent.com/anbulang/opencode-teleagent-auth/main/assets/icon.svg", "providers": ["teleagent"], - "summary": { "en": "Unofficial TeleAgent (天翼 AI 星辰超级智能体) desktop app models in magpie: the app's own sign-in, no API key, macOS and Windows, with points balance and daily check-in.", "zh": "非官方接入 TeleAgent(天翼 AI 星辰超级智能体)桌面端:复用桌面端登录态,不需要 API Key,支持 macOS 和 Windows,带积分余额与每日签到。" } }, - { "package": "opencode-xfyun-maas-auth", "name": "讯飞星辰 MaaS", "icon": "https://raw.githubusercontent.com/linjicong/opencode-xfyun-maas-auth/main/icon.svg", "providers": ["xfyun-maas", "xfyun-coding", "xfyun-token"], - "summary": { "en": "iFlytek Xfyun MaaS: Spark, GLM, DeepSeek, Kimi, Qwen and more — pay-as-you-go, Coding Plan and Token Plan.", "zh": "讯飞星辰 MaaS:星火、GLM、DeepSeek、Kimi、Qwen 等模型,按量付费与 Coding/Token 套餐。" } }, - { "package": "@magpie-community/page-usage-board", "name": "Usage Board", "kind": "page", "community": true, - "summary": { "en": "A quota card wall with expiry tiers, daily/per-provider charts, 24h cache-hit and token curves, optional remote attribution and account detail. Runs its own zero-dependency local service reading this machine's Magpie usage data; optional remote attribution queries the configured remote magpie.", "zh": "额度到期卡片墙 + 每日与 provider 用量图表 + 24 小时缓存命中与 token 曲线,可选远端归因与账号明细。自带零依赖本机服务,读本机 Magpie 用量数据;远端归因会查询所配置的远端 magpie。" } }, - { "package": "magpie-muse-auth", "name": "Muse (Meta)", "icon": "https://raw.githubusercontent.com/sonvivu/magpie-muse-auth/main/icon.svg", "providers": ["muse"], - "summary": { "en": "Muse (Meta) subscription for Magpie and OpenCode: device-code sign-in, minted API key, Wonder and Spark models with reasoning efforts. The -contributor models use Meta's Contributor lane (data may be used for training). Unofficial.", "zh": "使用 Muse (Meta) 订阅:设备码登录,自动换取 API 密钥,提供 Wonder 和 Spark 模型及推理强度设置;以 -contributor 结尾的模型走 Meta 的 Contributor 通道(数据可能用于训练)。(非官方)" } }, - { "package": "@magpie-community/page-model-quota", "name": "Model Quota", "kind": "page", "community": true, - "summary": { "en": "How many tokens each model has left, merged across every account and provider: each quota window is turned into tokens from your own usage history and split by the model's share in it.", "zh": "每个模型在所有账号、所有供应商的池子里还剩多少 token:用你自己的用量历史把每个额度窗口换算成 token,再按模型在窗口里的占比摊开。" } }, - { "package": "@magpie-community/page-harnessweight", "name": "HarnessWeight", "kind": "page", "community": true, - "summary": { "en": "Compare agent harnesses behind your gateway by input tokens, cache, failures, latency and estimated tool-definition size. Its local service reads this machine's usage ledger; a read-only hook samples tool sizes and tracks in-flight requests.", "zh": "按 Agent 对比网关后面的输入 token、缓存、失败率、延迟与工具定义体积估算。本机服务读取用量记录,只读 hook 采样工具大小并追踪进行中的请求。" } }, - { "package": "@magpie-community/agent-wave", "name": "Wave", "kind": "agent", "community": true, "icon": "https://github.com/wavetermdev.png", "summary": {"en": "Wave Terminal (waveterm) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put its config back.", "zh": "把 Wave Terminal(waveterm)加到 magpie 的 Agent 页:给它选 magpie 的模型,关掉即恢复它原来的配置。"}}, - { "package": "@magpie-community/agent-zycode", "name": "zycode", "kind": "agent", "community": true, "summary": {"en": "zycode (zypi, the VS Code extension, the ACP agent) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put agent.toml back. A host part copies the model into the [model] table of agent.toml, because magpie cannot edit TOML. Restart zycode after a change.", "zh": "把 zycode(zypi、VS Code 扩展、ACP agent)加到 magpie 的 Agent 页:给它选 magpie 的模型,关掉即恢复 agent.toml。magpie 不能编辑 TOML,所以由宿主部分把模型写进 agent.toml 的 [model] 表。切换后需重启 zycode。"}}, - { "package": "@magpie-community/agent-meowcode", "name": "MeowCode", "kind": "agent", "community": true, "summary": {"en": "MeowCode on magpie's Agents page: magpie's models appear in MeowCode's own model list, written to an overlay file MeowCode reads at start-up. Restart MeowCode after choosing a model.", "zh": "把 MeowCode 加到 magpie 的 Agent 页:magpie 的模型出现在 MeowCode 自己的模型列表里(启动时读取的覆盖层文件)。选择模型后需重启 MeowCode。"}}, - { "package": "@magpie-community/agent-lorca", "name": "Lorca", "kind": "agent", "community": true, "icon": "https://lorca.app/icon.png", "summary": {"en": "Lorca on magpie's Agents page: writes Magpie's gateway, models and MCP servers into Lorca's credentials (~/.lorca/credentials.json). Restart Lorca after connecting.", "zh": "把 Lorca 加到 magpie 的 Agent 页:将 Magpie 的网关、模型目录与 MCP 工具库写入 Lorca 的凭据配置(~/.lorca/credentials.json)。连接后需重启 Lorca。"}}, - { "package": "magpie-dafeiyu", "name": "DaFeiYu Desktop Pet", "kind": "page", - "summary": { "en": "A guide page for DaFeiYu, a desktop pet that lights a bulb per request and shows quota percentages and API balances. The pet is a separate Windows/macOS app you download yourself; this page does not install or start it. Artwork from PC2005-cloud/dsh-pet, non-commercial use only.", "zh": "蓝色大肥鱼桌宠的使用指南页:几个请求亮几盏灯,订阅显示额度百分比,API 钱包显示余额。桌宠是需要另外下载的 Windows/macOS 独立 App,本页不会安装或启动它。角色素材来自 PC2005-cloud/dsh-pet,仅限非商用。" } }, - { "package": "@magpie-community/agent-pine", "name": "Pine", "kind": "agent", "community": true, + { + "package": "magpie-deepaicode", + "name": "DeepAI Code", + "icon": "https://raw.githubusercontent.com/aqiu817/magpie-deepaicode/main/icon.svg", + "providers": [ + "deepaicode" + ], + "summary": { + "en": "Unofficial DeepAI Code access for magpie: browser sign-in, the desktop app's saved credentials or an API key, direct upstream requests with the device's real machine fingerprint, live models, daily Theos quota and check-in.", + "zh": "非官方接入 DeepAI Code:浏览器登录、桌面端凭据或 API Key,直连上游并使用真实设备指纹,动态模型目录、每日 Theos 额度与签到。" + } + }, + { + "package": "@magpie-community/opencode-agy-cli-auth", + "name": "Antigravity (local agy CLI)", + "icon": "antigravity-color", + "providers": [ + "agy-cli" + ], + "community": true, + "summary": { + "en": "Unofficial provider that runs the agy CLI you installed and signed in to through the Agent Client Protocol (agy acp), serving only the models agy offers this client. Unlike Antigravity ACP, which downloads Google's ACP server and rides the Antigravity IDE's sign-in, it uses your own agy and adds a process pool, a history budget, a bridge for the client's tools, session resume and agy's usage windows.", + "zh": "非官方插件:通过 Agent Client Protocol(agy acp)驱动你自己安装并登录的 agy CLI,只提供 agy 向本客户端开放的模型。与自行下载 Google ACP 服务端、依赖 Antigravity IDE 登录的 Antigravity ACP 不同,它用你本机的 agy,并带进程池、历史体积预算、客户端工具桥接、会话续接与 agy 用量窗口。" + } + }, + { + "package": "opencode-teleagent-auth", + "name": "TeleAgent 模型", + "icon": "https://raw.githubusercontent.com/anbulang/opencode-teleagent-auth/main/assets/icon.svg", + "providers": [ + "teleagent" + ], + "summary": { + "en": "Unofficial TeleAgent (天翼 AI 星辰超级智能体) desktop app models in magpie: the app's own sign-in, no API key, macOS and Windows, with points balance and daily check-in.", + "zh": "非官方接入 TeleAgent(天翼 AI 星辰超级智能体)桌面端:复用桌面端登录态,不需要 API Key,支持 macOS 和 Windows,带积分余额与每日签到。" + } + }, + { + "package": "opencode-xfyun-maas-auth", + "name": "讯飞星辰 MaaS", + "icon": "https://raw.githubusercontent.com/linjicong/opencode-xfyun-maas-auth/main/icon.svg", + "providers": [ + "xfyun-maas", + "xfyun-coding", + "xfyun-token" + ], + "summary": { + "en": "iFlytek Xfyun MaaS: Spark, GLM, DeepSeek, Kimi, Qwen and more — pay-as-you-go, Coding Plan and Token Plan.", + "zh": "讯飞星辰 MaaS:星火、GLM、DeepSeek、Kimi、Qwen 等模型,按量付费与 Coding/Token 套餐。" + } + }, + { + "package": "@magpie-community/page-usage-board", + "name": "Usage Board", + "kind": "page", + "community": true, + "summary": { + "en": "A quota card wall with expiry tiers, daily/per-provider charts, 24h cache-hit and token curves, optional remote attribution and account detail. Runs its own zero-dependency local service reading this machine's Magpie usage data; optional remote attribution queries the configured remote magpie.", + "zh": "额度到期卡片墙 + 每日与 provider 用量图表 + 24 小时缓存命中与 token 曲线,可选远端归因与账号明细。自带零依赖本机服务,读本机 Magpie 用量数据;远端归因会查询所配置的远端 magpie。" + } + }, + { + "package": "magpie-muse-auth", + "name": "Muse (Meta)", + "icon": "https://raw.githubusercontent.com/sonvivu/magpie-muse-auth/main/icon.svg", + "providers": [ + "muse" + ], + "summary": { + "en": "Muse (Meta) subscription for Magpie and OpenCode: device-code sign-in, minted API key, Wonder and Spark models with reasoning efforts. The -contributor models use Meta's Contributor lane (data may be used for training). Unofficial.", + "zh": "使用 Muse (Meta) 订阅:设备码登录,自动换取 API 密钥,提供 Wonder 和 Spark 模型及推理强度设置;以 -contributor 结尾的模型走 Meta 的 Contributor 通道(数据可能用于训练)。(非官方)" + } + }, + { + "package": "@magpie-community/page-model-quota", + "name": "Model Quota", + "kind": "page", + "community": true, + "summary": { + "en": "How many tokens each model has left, merged across every account and provider: each quota window is turned into tokens from your own usage history and split by the model's share in it.", + "zh": "每个模型在所有账号、所有供应商的池子里还剩多少 token:用你自己的用量历史把每个额度窗口换算成 token,再按模型在窗口里的占比摊开。" + } + }, + { + "package": "@magpie-community/page-harnessweight", + "name": "HarnessWeight", + "kind": "page", + "community": true, + "summary": { + "en": "Compare agent harnesses behind your gateway by input tokens, cache, failures, latency and estimated tool-definition size. Its local service reads this machine's usage ledger; a read-only hook samples tool sizes and tracks in-flight requests.", + "zh": "按 Agent 对比网关后面的输入 token、缓存、失败率、延迟与工具定义体积估算。本机服务读取用量记录,只读 hook 采样工具大小并追踪进行中的请求。" + } + }, + { + "package": "@magpie-community/agent-wave", + "name": "Wave", + "kind": "agent", + "community": true, + "icon": "https://github.com/wavetermdev.png", + "summary": { + "en": "Wave Terminal (waveterm) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put its config back.", + "zh": "把 Wave Terminal(waveterm)加到 magpie 的 Agent 页:给它选 magpie 的模型,关掉即恢复它原来的配置。" + } + }, + { + "package": "@magpie-community/agent-zycode", + "name": "zycode", + "kind": "agent", + "community": true, + "summary": { + "en": "zycode (zypi, the VS Code extension, the ACP agent) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put agent.toml back. A host part copies the model into the [model] table of agent.toml, because magpie cannot edit TOML. Restart zycode after a change.", + "zh": "把 zycode(zypi、VS Code 扩展、ACP agent)加到 magpie 的 Agent 页:给它选 magpie 的模型,关掉即恢复 agent.toml。magpie 不能编辑 TOML,所以由宿主部分把模型写进 agent.toml 的 [model] 表。切换后需重启 zycode。" + } + }, + { + "package": "@magpie-community/agent-meowcode", + "name": "MeowCode", + "kind": "agent", + "community": true, + "summary": { + "en": "MeowCode on magpie's Agents page: magpie's models appear in MeowCode's own model list, written to an overlay file MeowCode reads at start-up. Restart MeowCode after choosing a model.", + "zh": "把 MeowCode 加到 magpie 的 Agent 页:magpie 的模型出现在 MeowCode 自己的模型列表里(启动时读取的覆盖层文件)。选择模型后需重启 MeowCode。" + } + }, + { + "package": "@magpie-community/agent-lorca", + "name": "Lorca", + "kind": "agent", + "community": true, + "icon": "https://lorca.app/icon.png", + "summary": { + "en": "Lorca on magpie's Agents page: writes Magpie's gateway, models and MCP servers into Lorca's credentials (~/.lorca/credentials.json). Restart Lorca after connecting.", + "zh": "把 Lorca 加到 magpie 的 Agent 页:将 Magpie 的网关、模型目录与 MCP 工具库写入 Lorca 的凭据配置(~/.lorca/credentials.json)。连接后需重启 Lorca。" + } + }, + { + "package": "magpie-dafeiyu", + "name": "DaFeiYu Desktop Pet", + "kind": "page", + "summary": { + "en": "A guide page for DaFeiYu, a desktop pet that lights a bulb per request and shows quota percentages and API balances. The pet is a separate Windows/macOS app you download yourself; this page does not install or start it. Artwork from PC2005-cloud/dsh-pet, non-commercial use only.", + "zh": "蓝色大肥鱼桌宠的使用指南页:几个请求亮几盏灯,订阅显示额度百分比,API 钱包显示余额。桌宠是需要另外下载的 Windows/macOS 独立 App,本页不会安装或启动它。角色素材来自 PC2005-cloud/dsh-pet,仅限非商用。" + } + }, + { + "package": "@magpie-community/agent-pine", + "name": "Pine", + "kind": "agent", + "community": true, "icon": "https://raw.githubusercontent.com/pmndre/pine/next/docs/images/logo-light.svg", - "summary": { "en": "Pine (pmndre/pine, the event-driven agent harness on pi's SDK) on magpie's Agents page: pick one of magpie's models and magpie writes its gateway and catalog into ~/.pine/models.json and the model into pine.json, so new conversations start on it. Restart Pine after a change; unofficial.", "zh": "把 Pine(pmndre/pine,跑在 pi SDK 上的事件驱动 Agent 框架)加到 magpie 的 Agent 页:选一个 magpie 的模型,magpie 把网关和模型目录写进 ~/.pine/models.json、把起始模型写进 pine.json,新对话就用它。换完模型要重启 Pine;非官方适配。" } } + "summary": { + "en": "Pine (pmndre/pine, the event-driven agent harness on pi's SDK) on magpie's Agents page: pick one of magpie's models and magpie writes its gateway and catalog into ~/.pine/models.json and the model into pine.json, so new conversations start on it. Restart Pine after a change; unofficial.", + "zh": "把 Pine(pmndre/pine,跑在 pi SDK 上的事件驱动 Agent 框架)加到 magpie 的 Agent 页:选一个 magpie 的模型,magpie 把网关和模型目录写进 ~/.pine/models.json、把起始模型写进 pine.json,新对话就用它。换完模型要重启 Pine;非官方适配。" + } + } ] } From ad835bc641fe8eff51a78a04a2c749d586add089 Mon Sep 17 00:00:00 2001 From: chenyan Date: Wed, 18 Nov 2026 20:20:14 +0800 Subject: [PATCH 2/2] registry: one compact entry for openrouter-decisions --- registry.json | 1608 ++++++++----------------------------------------- 1 file changed, 236 insertions(+), 1372 deletions(-) diff --git a/registry.json b/registry.json index c72255ec8..3e5801c2e 100644 --- a/registry.json +++ b/registry.json @@ -1,1131 +1,205 @@ { "version": 1, "plugins": [ - { - "package": "@magpie-community/opencode-gumloop-auth", - "name": "Gumloop", - "icon": "https://www.gumloop.com/favicon.ico", - "providers": [ - "gumloop" - ], - "community": true, - "summary": { - "en": "Gumloop with your API key and user ID: account models, streaming, tool calls and a locally calibrated credit estimate.", - "zh": "使用 API key 和用户 ID 接入 Gumloop:按账号发现模型,支持流式响应、工具调用和本地校准的积分估算。" - } - }, - { - "package": "@magpie-community/opencode-agnes-auth", - "name": "Agnes AI", - "icon": "https://www.agnes-ai.com/images/biglogo.png", - "providers": [ - "agnes" - ], - "community": true, - "summary": { - "en": "Agnes AI international with Chrome browser sign-in: live OpenAI-compatible text models, Token Plan and Agent Plan usage, and the USD wallet balance.", - "zh": "通过 Chrome 浏览器登录 Agnes AI 国际版:动态发现 OpenAI 兼容文本模型,显示 Token Plan、Agent Plan 用量与美元余额,无需手动复制会话 token。" - } - }, - { - "package": "@magpie-community/opencode-atomcode-auth", - "name": "AtomCode", - "icon": "https://atomgit.com/favicon.ico", - "providers": [ - "atomcode" - ], - "community": true, - "summary": { - "en": "Unofficial AtomGit CodingPlan access for magpie and OpenCode: browser OAuth or local AtomCode CLI sign-in, signed OpenAI-compatible requests, live models, quota windows, and a read-only diagnostics page.", - "zh": "非官方接入 AtomGit CodingPlan:支持浏览器 OAuth 或复用本机 AtomCode CLI 登录,发送签名请求,显示实时模型与额度窗口,并提供只读诊断页。" - } - }, - { - "package": "@magpie-community/gateway-powercontext", - "name": "PowerContext", - "kind": "middleware", - "community": true, - "icon": "https://raw.githubusercontent.com/oceanbase/powercontext/master/src/powercontext/server/dashboard/static/brand/favicon-color.svg", - "summary": { - "en": "Recall from your PowerContext service and optionally save complete conversations. Configure the service URL first; capture is off by default.", - "zh": "从自己的 PowerContext 服务召回记忆,并按需存回完整对话。先配置服务地址;对话存回默认关闭。" - } - }, - { - "package": "@magpie-community/gateway-nowledge-mem", - "name": "Nowledge Mem", - "kind": "middleware", - "community": true, - "summary": { - "en": "Give every harness behind Magpie Nowledge Mem automatically: recall and the Context Bundle (or Working Memory) before each request, per-harness provenance, optional import of completed conversations as Mem Threads, and a real connection test in the page. A local or remote Mem; saving conversations is off by default.", - "zh": "让 Magpie 后面的每个 harness 自动接入 Nowledge Mem:每次请求前召回记忆与 Context Bundle(缺失时用 Working Memory),记录来自哪个 harness,按需把已完成的对话导入为 Mem Thread,并在页面中提供真实的连接测试。本机或远端 Mem 都可用;对话导入默认关闭。" - } - }, - { - "package": "@magpie-community/gateway-aelios", - "name": "Aelios", - "kind": "middleware", - "community": true, - "summary": { - "en": "Long-term memory from your own Aelios on Cloudflare: recall before each turn for every agent behind magpie, and optionally record the conversation so Aelios can learn it. Set it up on the plugin's own page, with a connection test and per-model or per-agent routing to separate assistants. Recording is off by default.", - "zh": "接入你自己部署在 Cloudflare 上的 Aelios 长期记忆:magpie 后面的每个 agent 每轮自动召回,按需把对话交给 Aelios 记住。在插件自己的页面里配置,可选择连接测试,并按模型或客户端分流到不同的助手。记录默认关闭。" - } - }, - { - "package": "@magpie-community/gateway-respire-memory", - "name": "respire-memory", - "kind": "middleware", - "community": true, - "icon": "https://raw.githubusercontent.com/risense-ai/respire-magpie/v0.3.0/page/icon.svg", - "summary": { - "en": "Help AI agents evolve through experience: per-turn recall and complete engine-policy reads, automatic process diaries and grounded archival with write readback. Configure an existing local Runtime and generation model first. Archival defaults on; raw capture off. Background storage happens after reply completion.", - "zh": "让 AI Agent 通过经验积累持续自进化:每轮召回并读完整政策,自动保存过程日记与整理长期记忆,写后回读。先配置已有本地 Runtime 和生成模型;归档默认开启,原始采集关闭。宿主限制下存储在回复完成后执行。" - } - }, - { - "package": "@openviking/magpie-plugin", - "name": "OpenViking", - "kind": "middleware", - "community": true, - "icon": "https://raw.githubusercontent.com/volcengine/OpenViking/main/docs/images/ov-logo-icon.png", - "summary": { - "en": "Recall OpenViking context before model requests and optionally capture completed conversations. Reuses your existing OpenViking configuration; capture is off by default. Maintained by the OpenViking project.", - "zh": "在模型请求前召回 OpenViking 上下文,并按需采集完整对话。复用现有 OpenViking 配置;采集默认关闭。由 OpenViking 项目维护。" - } - }, - { - "package": "@vshulcz/magpie-deja", - "name": "deja-vu", - "kind": "middleware", - "community": true, - "icon": "https://raw.githubusercontent.com/vshulcz/deja-vu/main/docs/assets/icon.svg", - "summary": { - "en": "Recall local session history from the transcripts Claude Code, Codex, Hermes, Gemini CLI, OpenCode, Cursor and DSH write on this machine before each request. Scopes lookup to the request's project directory (ctx.cwd) by default and scrubs secret patterns before attaching recalled excerpts to the request sent to your model provider. When recalled history helped, the model is asked to end its reply with a one-line \"déjà vu: … — reusing it.\" note.", - "zh": "每次请求前从本机 Claude Code、Codex、Hermes、Gemini CLI、OpenCode、Cursor 与 DSH 的磁盘会话记录中召回相关历史。默认严格限定在请求所属的项目目录(ctx.cwd)内检索,并对密钥与令牌执行脱敏后,将召回片段随请求一并发送给所配置的模型供应商。当召回的历史记录有帮助时,会请模型在回复末尾附上一行“déjà vu: … — reusing it.”说明。" - } - }, - { - "package": "@magpie-community/opencode-copilot-auth", - "name": "GitHub Copilot", - "icon": "githubcopilot", - "providers": [ - "github-copilot" - ], - "community": true, - "summary": { - "en": "GitHub Copilot on GitHub.com or GitHub Enterprise, with premium request or credit usage, remaining allowance and quota reset dates.", - "zh": "GitHub Copilot:支持 GitHub.com 与 GitHub Enterprise,显示高级请求或积分用量、剩余额度和重置时间。" - } - }, - { - "package": "@magpie-community/opencode-joycode-auth", - "name": "JoyCode", - "providers": [ - "joycode" - ], - "community": true, - "summary": { - "en": "Unofficial JoyCode provider: JD account browser sign-in, live models and credit multipliers, plan credits and streaming tool calls.", - "zh": "非官方 JoyCode 插件:京东账号浏览器登录,动态模型与积分倍率、套餐额度及流式工具调用。" - } - }, - { - "package": "@magpie-community/opencode-doubao-work-auth", - "name": "豆包工作", - "icon": "doubao-color", - "providers": [ - "doubao-work" - ], - "community": true, - "summary": { - "en": "Not a model API: Doubao Work has none for other clients. This drives its desktop app (or Doubao's) the way a person does, so each request is one run of Doubao's cloud agent in a new task, with the client's tools asked for as text. macOS and Windows; unofficial.", - "zh": "不是模型接口:豆包工作没有给第三方客户端用的模型接口。本插件像人一样操作豆包工作(或豆包)桌面版,每个请求是豆包云端智能体在一个新任务里的一次执行,客户端的工具以文本方式调用。macOS 和 Windows;非官方。" - } - }, - { - "package": "@magpie-community/opencode-kimi-work-auth", - "name": "Kimi Work", - "icon": "kimi", - "providers": [ - "kimi-work" - ], - "community": true, - "summary": { - "en": "Kimi desktop Work subscription through its local sign-in, with live Work models and shared membership usage. Separate from Kimi Code.", - "zh": "使用本机普通 Kimi 桌面版的 Work 登录,自动读取 Work 模型和会员共享用量;独立于 Kimi Code。" - } - }, - { - "package": "@magpie-community/opencode-zen-free-auth", - "name": "OpenCode Zen Free", - "icon": "opencode", - "providers": [ - "opencode-zen-free" - ], - "community": true, - "summary": { - "en": "OpenCode Zen's free models with the public credential, automatic discovery and native streaming protocols.", - "zh": "使用 public 凭证接入 OpenCode Zen 免费模型,自动发现模型并支持原生流式协议。" - } - }, - { - "package": "@magpie-community/opencode-zcode-auth", - "name": "ZCode", - "icon": "zcode", - "providers": [ - "zcode" - ], - "community": true, - "replaces": "zcode", - "summary": { - "en": "Z.ai and BigModel GLM Coding Plan, team seats and ZCode's free Start Plan, with a daily claim of its gift plans.", - "zh": "Z.ai 与智谱 GLM Coding Plan,支持团队席位和 ZCode 免费 Start Plan,可每日领取 ZCode 活动包。" - } - }, - { - "package": "@magpie-community/opencode-qoder-auth", - "name": "Qoder", - "icon": "qoder", - "providers": [ - "qoder", - "qoder-cn", - "qwenwork", - "qwenwork-enterprise" - ], - "community": true, - "replaces": "qoder", - "summary": { - "en": "Your Qoder subscription: international (qoder.com) or Qoder CN, the China site (qoder.cn, Alibaba Cloud or phone sign-in), each its own provider. Also QwenWork (千问办公): the public cloud or the enterprise flagship with its own domain. Unofficial for QwenWork.", - "zh": "用 Qoder 自己的设备登录使用 Qoder 订阅:国际版(qoder.com)与 Qoder 中国版(Qoder CN,qoder.cn,阿里云账号 / 手机号登录)各是一个订阅;也接入千问办公:公有云与企业旗舰版(输入企业自己的域名)。千问办公部分非官方。" - } - }, - { - "package": "@magpie-community/opencode-devin-auth", - "name": "Devin", - "icon": "devin", - "providers": [ - "devin" - ], - "community": true, - "replaces": "devin", - "summary": { - "en": "Devin's models with your Devin account — a browser sign-in or the devin CLI's.", - "zh": "用 Devin 账号使用 Devin 的模型:浏览器登录,或沿用 devin CLI 的登录。" - } - }, - { - "package": "@magpie-community/opencode-zed-auth", - "name": "Zed", - "icon": "zed", - "providers": [ - "zed" - ], - "community": true, - "replaces": "zed", - "summary": { - "en": "Claude, GPT, Gemini and Grok hosted by Zed, on your Zed plan.", - "zh": "用 Zed 订阅使用 Zed 托管的 Claude、GPT、Gemini 和 Grok。" - } - }, - { - "package": "@magpie-community/opencode-factory-auth", - "name": "Factory", - "icon": "factory", - "providers": [ - "factory" - ], - "community": true, - "replaces": "factory", - "summary": { - "en": "Factory's Droid subscription: models from Anthropic, OpenAI, xAI and open labs.", - "zh": "Factory Droid 订阅:Anthropic、OpenAI、xAI 及开源模型。" - } - }, - { - "package": "@magpie-community/opencode-grok-auth", - "name": "Grok", - "icon": "xai", - "providers": [ - "grok" - ], - "community": true, - "replaces": "grok", - "summary": { - "en": "SuperGrok and X Premium+, through the Grok Build CLI's sign-in.", - "zh": "通过 Grok Build CLI 的登录使用 SuperGrok 和 X Premium+。" - } - }, - { - "package": "@magpie-community/opencode-commandcode-auth", - "name": "Command Code", - "icon": "commandcode", - "providers": [ - "commandcode-plan" - ], - "community": true, - "replaces": "commandcode-plan", - "summary": { - "en": "Command Code plans — Pro, GOAT, Max, Ultra, Go and Teams.", - "zh": "Command Code 各档套餐:Pro、GOAT、Max、Ultra、Go 和 Teams。" - } - }, - { - "package": "@magpie-community/opencode-mimo-auth", - "name": "Xiaomi MiMo", - "icon": "mimocode", - "providers": [ - "mimo-app" - ], - "community": true, - "replaces": "mimo-app", - "summary": { - "en": "Xiaomi MiMo models with your Xiaomi account.", - "zh": "用小米账号使用小米 MiMo 模型。" - } - }, - { - "package": "@magpie-community/opencode-cursor-auth", - "name": "Cursor", - "icon": "cursor", - "providers": [ - "cursor" - ], - "community": true, - "replaces": "cursor", - "summary": { - "en": "Cursor Pro, Pro+, Ultra and Teams — every model your plan has, with tool calls.", - "zh": "Cursor Pro、Pro+、Ultra 与 Teams:套餐里的所有模型,支持工具调用。" - } - }, - { - "package": "@magpie-community/opencode-kiro-auth", - "name": "Kiro", - "icon": "kiro-color", - "providers": [ - "kiro" - ], - "community": true, - "replaces": "kiro", - "summary": { - "en": "Kiro Free, Pro, Pro+ and Power — Kiro's sign-in, kiro-cli's, an API key, or accounts imported by refresh token.", - "zh": "Kiro Free、Pro、Pro+ 与 Power:Kiro 登录、沿用 kiro-cli 的登录、API key,或按 refresh token 导入账号。" - } - }, - { - "package": "@magpie-community/opencode-minimax-auth", - "name": "MiniMax Code", - "icon": "minimax-color", - "providers": [ - "minimax-code", - "minimax-code-global" - ], - "community": true, - "summary": { - "en": "MiniMax Code credits and M Plan with device sign-in, native web search, image/video generation and media tools through Magpie MCP.", - "zh": "MiniMax Code 积分与 M Plan、设备登录,提供原生网络搜索、生图、生视频及多模态 MCP 工具。" - } - }, - { - "package": "@magpie-community/opencode-warp-auth", - "name": "Warp", - "icon": "warp", - "providers": [ - "warp" - ], - "community": true, - "summary": { - "en": "Warp (the terminal) AI request credits through the Warp app's sign-in: the auto routers and every model family the account lists.", - "zh": "用 Warp 终端应用的登录使用 Warp AI 请求额度:auto 路由与账户可用的全部模型家族。" - } - }, - { - "package": "@magpie-community/opencode-workbuddy-auth", - "name": "WorkBuddy", - "icon": "workbuddy-color", - "providers": [ - "workbuddy", - "workbuddy-ai" - ], - "community": true, - "replaces": "workbuddy", - "summary": { - "en": "WorkBuddy's China build (CodeBuddy plan) and its international build.", - "zh": "WorkBuddy 国内版(CodeBuddy 套餐)与国际版。" - } - }, - { - "package": "@magpie-community/opencode-cline-auth", - "name": "Cline", - "icon": "cline", - "providers": [ - "cline" - ], - "community": true, - "summary": { - "en": "Cline's models on your own Cline account — usage billing, ClinePass and Cline's free models — with Cline's device sign-in or an API key.", - "zh": "用自己的 Cline 账号使用 Cline 的模型——用量计费、ClinePass 与免费模型——支持 Cline 设备登录或 API key。" - } - }, - { - "package": "@magpie-community/opencode-trae-auth", - "name": "Trae", - "icon": "trae", - "providers": [ - "trae-cn", - "trae-global" - ], - "community": true, - "summary": { - "en": "Trae CN (trae.cn) and Trae international (trae.ai) models with your own account, free tier included — Trae's own browser sign-in. Experimental.", - "zh": "用 Trae CN(trae.cn)与 Trae 国际版(trae.ai)账号使用各自的模型,含免费版,使用 Trae 自己的浏览器登录。实验性。" - } - }, - { - "package": "@magpie-community/opencode-opendesign-auth", - "name": "Open Design", - "providers": [ - "opendesign" - ], - "community": true, - "summary": { - "en": "Open Design Cloud's Design Plan (open-design.ai): the account's device sign-in, the models the plan entitles it to, and the plan's 5-hour and 7-day allowance and wallet balance.", - "zh": "Open Design Cloud 的 Design Plan(open-design.ai):账号设备登录、套餐可用的模型,以及 5 小时与 7 天额度、钱包余额。" - } - }, - { - "package": "@magpie-community/opencode-catpaw-auth", - "name": "CatPaw", - "providers": [ - "catpaw" - ], - "community": true, - "summary": { - "en": "Meituan's 妙手 CatPaw (ai.catpaw.meituan.com): a browser web sign-in or the desktop client's saved credentials, chat through its round/event/turn conversation protocol, models and credit balance.", - "zh": "美团妙手 CatPaw(ai.catpaw.meituan.com):浏览器网页登录或妙手桌面端已保存的登录态,经其 round/event/turn 会话协议对话,含模型与积分余额。" - } - }, - { - "package": "@magpie-community/opencode-loomy-auth", - "name": "Loomy", - "providers": [ - "loomy" - ], - "community": true, - "summary": { - "en": "iFlytek's Loomy (loomyad.xunfei.cn): SMS sign-in, OpenAI-compatible chat with reasoning-effort folding and image markers, the live model catalog, two point wallets and the daily first-login grant.", - "zh": "讯飞 Loomy(loomyad.xunfei.cn):短信验证码登录,OpenAI 兼容对话(思考档折叠与图片标记适配),实时模型目录、双积分钱包与每日首登赠送。" - } - }, - { - "package": "@magpie-community/opencode-nous-auth", - "name": "Nous Portal (OAuth)", - "icon": "nousresearch", - "providers": [ - "nous-portal" - ], - "community": true, - "summary": { - "en": "Nous Portal models, with Nous Portal's device sign-in.", - "zh": "通过 Nous Portal 自己的设备码登录使用 Nous Portal 模型。" - } - }, - { - "package": "@magpie-community/opencode-raccoon-auth", - "name": "商汤小浣熊", - "icon": "sensenova-color", - "providers": [ - "raccoon" - ], - "community": true, - "summary": { - "en": "SenseTime Raccoon (xiaohuanxiong.com) models, account credits and daily check-in, with browser sign-in or Raccoon desktop's live sign-in.", - "zh": "商汤小浣熊(xiaohuanxiong.com)模型、积分余额与每日签到,支持浏览器登录或直接复用小浣熊桌面端登录态。" - } - }, - { - "package": "@magpie-community/opencode-sensenova-auth", - "name": "SenseNova", - "icon": "sensenova-color", - "providers": [ - "sensenova-token-plan" - ], - "community": true, - "summary": { - "en": "SenseNova (SenseTime 日日新) Token Plan: a browser sign-in (or the console account and password) reads the credit pools, their 5-hour and 7-day windows and the expiring grants, and serves the plan's own models — the browser sign-in reuses or creates a Token Plan API key for you.", - "zh": "商汤日日新 Token Plan:浏览器登录(或控制台账号密码)读取积分池、5 小时 / 7 天窗口与将过期的返赠余额,并直接调用计划内模型——浏览器登录会自动复用或创建 Token Plan API Key。" - } - }, - { - "package": "@magpie-community/opencode-inkstone-auth", - "name": "书生·端砚", - "icon": "https://raw.githubusercontent.com/magpie-community/plugins/main/packages/inkstone/icon.svg", - "providers": [ - "intern-inkstone" - ], - "community": true, - "summary": { - "en": "Unofficial: Intern InkStone(书生·端砚) Token Plan: a browser sign-in reads the ink-credit balance, the 5-hour and 7-day windows and the credit packs, and creates an ink-* API key (or reuses a pasted one) to serve the plan's models for chat.", - "zh": "非官方:书生·端砚(Intern InkStone)墨点计划:浏览器登录读取墨点余额、5 小时 / 7 天窗口与墨点包,并创建 ink-* API Key(或复用现有 Key)直接调用计划内模型对话。" - } - }, - { - "package": "@magpie-community/opencode-autoclaw-auth", - "name": "AutoClaw", - "providers": [ - "autoclaw", - "autoclaw-cn" - ], - "community": true, - "summary": { - "en": "AutoClaw (Zhipu OpenClaw) cloud accounts: oversea web sign-in (Google/z.ai) or CN phone SMS, credit balance and daily check-in.", - "zh": "AutoClaw(智谱澳龙/OpenClaw)云端账号:国际版网页登录(Google/z.ai)或国内手机验证码,积分余额与每日签到。" - } - }, - { - "package": "@magpie-community/opencode-codearts-auth", - "name": "CodeArts / OfficeAce", - "providers": [ - "codearts", - "officeace" - ], - "community": true, - "summary": { - "en": "Unofficial: Huawei Cloud CodeArts (码道) coding plans and OfficeAce (AgentArts), each signed in with Huawei Cloud OAuth on its own card — GLM, DeepSeek, openPangu and OfficeAce's MaaS models, live models and credit rates, credit usage by bucket and the daily check-in. Replaces officeace-auth; accounts signed in there keep working.", - "zh": "非官方:华为云码道(CodeArts)编码套餐与 OfficeAce(AgentArts),两张卡各自用华为云 OAuth 登录;GLM、DeepSeek、openPangu 与 OfficeAce 的 MaaS 模型,实时模型与积分倍率、分桶积分用量和每日签到。取代 officeace-auth,原已登录的账号继续可用。" - } - }, - { - "package": "@magpie-community/opencode-gcp-ai-auth", - "name": "GCP AI", - "providers": [ - "gcp-ai" - ], - "community": true, - "summary": { - "en": "Vertex AI models and Google Cloud promotional credits, signed with your own gcloud or ADC login. Lists only the models your project can call.", - "zh": "用本机 gcloud / ADC 登录使用 Vertex AI 模型,并读取 Google Cloud 促销赠金;只列出项目实际能调用的模型。" - } - }, - { - "package": "magpie-review-rules", - "name": "Claude Desktop Review & Model Rules", - "kind": "middleware", - "summary": { - "en": "Pin Claude Desktop auto-mode action reviews to a chosen Magpie model, independently of the conversation model. Configure per-model Claude compatibility presets and sandboxed JavaScript request rules in a native settings page. No request changes until configured.", - "zh": "固定 Claude Desktop 自动模式的操作审核模型,不随对话模型切换;通过原生设置页配置逐模型 Claude 兼容预设与沙箱 JavaScript 请求规则。首次安装不改写请求。" - } - }, - { - "package": "@magpie-community/middleware-param-override", - "name": "Param Override", - "kind": "middleware", - "community": true, - "summary": { - "en": "New API's parameter override: set, delete, move or rewrite fields of every request, by model or by condition.", - "zh": "New API 的参数覆盖:按模型或条件设置、删除、移动或改写每个请求的字段。" - } - }, - { - "package": "@magpie-community/middleware-model-map", - "name": "Model Map", - "kind": "middleware", - "community": true, - "summary": { - "en": "New API's model redirection: send a model under another name, by name or pattern; replies keep the name asked for.", - "zh": "New API 的模型重定向:按名称或正则把模型换成另一个发出,回复里仍是原来的名字。" - } - }, - { - "package": "@magpie-community/middleware-system-prompt", - "name": "System Prompt", - "kind": "middleware", - "community": true, - "summary": { - "en": "Add your own system prompt to every request, or only some agents' or models', in each API's own place.", - "zh": "给每个请求(或指定 agent、模型)加上自己的系统提示词,放在各 API 自己的位置。" - } - }, - { - "package": "@magpie-community/middleware-output-strategy", - "name": "Output Strategy", - "kind": "middleware", - "community": true, - "summary": { - "en": "Seven composable output styles, all off by default, with a selection panel and prompt-cache-stable injection that preserves Codex instructions. Unofficial; by manjughosa.", - "zh": "七种可组合的输出样式,全部默认关闭;带选择面板,注入不破坏提示缓存,也不改写 Codex instructions。由 manjughosa 独立开发并维护,非官方出品。" - } - }, - { - "package": "@magpie-community/middleware-word-guard", - "name": "Word Guard", - "kind": "middleware", - "community": true, - "summary": { - "en": "New API's sensitive-word filter: turn away or mask words and patterns in what you send, and in replies.", - "zh": "New API 的敏感词过滤:拒绝或替换你发出的内容和回复里的词与正则。" - } - }, - { - "package": "@magpie-community/middleware-company-guard", - "name": "Company Guard", - "kind": "middleware", - "community": true, - "summary": { - "en": "Recognize and mask company names with stable session aliases, restore replies, and show local receipts and explicit Codex session approvals. Unlike Word Guard, names can be restored. Text only; rule-based detection.", - "zh": "识别公司名称并使用会话内稳定代号,回复返回本机时还原,提供本地回执与需明确确认的 Codex 会话授权。区别于 Word Guard 的词表拒绝或替换,支持名称还原;仅处理文字,采用规则识别。" - } - }, - { - "package": "@magpie-community/middleware-think-tags", - "name": "Think Tags", - "kind": "middleware", - "community": true, - "summary": { - "en": "Take a model's leading , or block out of its reply's text, or put a Chat reply's reasoning into it.", - "zh": "去掉回复正文开头的 、 或 思考块,或把 Chat 回复的思考内容放进正文。" - } - }, - { - "package": "@magpie-community/middleware-empty-reply", - "name": "Empty Reply", - "kind": "middleware", - "community": true, - "summary": { - "en": "Drop the end of a reply that has reasoning but no text, so Claude Code and Codex retry instead of treating an empty turn as finished. Anthropic and Responses, streamed or whole.", - "zh": "丢掉「只有思考、没有正文」回复的结尾,让 Claude Code 和 Codex 重试,而不是把空回合当成正常结束。支持 Anthropic 与 Responses 的流式与整包回复。" - } - }, - { - "package": "@magpie-community/agent-nowledge-mem", - "name": "Nowledge Mem / AI Now models", - "kind": "agent", - "community": true, - "icon": "https://mem.nowledge.co/favicon.png", - "summary": { - "en": "Choose Nowledge Mem Default, AI Now and Background models through Magpie. Distinct from gateway-nowledge-mem recall/capture. Disconnect before uninstalling.", - "zh": "在 Magpie 管理 Nowledge 的 Default、AI Now 和后台模型;与 gateway-nowledge-mem 的记忆召回、对话导入用途不同。卸载前先断开。" - } - }, - { - "package": "@magpie-community/agent-kilo", - "name": "Kilo Code", - "kind": "agent", - "community": true, - "icon": "https://github.com/Kilo-Org.png", - "summary": { - "en": "Kilo Code (kilo.ai) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put its config back.", - "zh": "把 Kilo Code(kilo.ai)加到 magpie 的 Agent 页:选择 magpie 的模型,关掉即恢复原来的配置。" - } - }, - { - "package": "@magpie-community/middleware-antigravity-cloak", - "name": "Antigravity Cloak", - "kind": "middleware", - "community": true, - "summary": { - "en": "Claude Code and Claude Desktop on an Antigravity account: the words Antigravity refuses their system prompt for are broken up, in what Antigravity alone is sent.", - "zh": "让 Claude Code 和 Claude Desktop 能用 Antigravity 账号:只在发给 Antigravity 的请求里拆开它据以拒绝系统提示词的那几个词。" - } - }, - { - "package": "@magpie-community/middleware-fovea", - "name": "Fovea", - "kind": "middleware", - "community": true, - "summary": { - "en": "Thin an agent's old history while preserving a recent window and tool pairing. Optional saccade collapses old content on a new topic; a structured state block replaces the raw task message; cut boundaries move in blocks to keep provider prompt caches settling.", - "zh": "裁薄 Agent 旧历史,保留最近窗口与工具配对;可选 saccade 在新话题时折叠旧内容,结构化状态块替代原始任务消息,裁剪边界分块移动以保护 prompt cache。" - } - }, - { - "package": "@magpie-community/middleware-codex-history", - "name": "Codex History Filter", - "kind": "middleware", - "community": true, - "summary": { - "en": "Opt-in, per-provider Codex Responses history filtering after switching providers. Preserves readable content and tool-call pairing; no model probes.", - "zh": "按供应商开关 Codex Responses 历史过滤,缓解切换上游后的密文和历史 ID 兼容问题;保留正文与工具配对,不运行模型检测。" - } - }, - { - "package": "@magpie-community/middleware-history-trim", - "name": "History Trim", - "kind": "middleware", - "community": true, - "summary": { - "en": "Send a request's history smaller: fold a run of blank lines, drop a system prompt written twice and a message repeated right after itself, and cut a past tool result longer than maxToolLength at a word boundary. The turn being sent now is left whole.", - "zh": "把请求的历史发得更小:折叠连续空行、去掉逐字重复的 system 提示与紧挨着重复的消息、把历史中超过 maxToolLength 的工具结果在词边界处截断;正在发送的这一轮原样不动。" - } - }, - { - "package": "@magpie-community/agent-agy", - "name": "Antigravity CLI", - "kind": "agent", - "community": true, - "icon": "antigravity-color", - "summary": { - "en": "Antigravity CLI (agy) on magpie's Agents page: pick one of magpie's models, and the agy in your shells starts on magpie's gateway. Takes the place of the built-in, keeping what it set up.", - "zh": "把 Antigravity CLI(agy)加到 magpie 的 Agent 页:选一个 magpie 的模型,终端里的 agy 就从 magpie 的网关启动。接替内置的 agy,保留它原有的设置。" - } - }, - { - "package": "@magpie-community/agent-aider", - "name": "Aider", - "kind": "agent", - "community": true, - "summary": { - "en": "Aider (aider.chat) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put its config back.", - "zh": "把 Aider(aider.chat)加到 magpie 的 Agent 页:给它选 magpie 的模型,关掉即恢复它原来的配置。" - } - }, - { - "package": "@magpie-community/agent-nanobot", - "name": "nanobot", - "kind": "agent", - "community": true, - "icon": "https://raw.githubusercontent.com/HKUDS/nanobot/main/webui/public/brand/nanobot_mark.svg", - "summary": { - "en": "nanobot (HKUDS/nanobot) on magpie's Agents page: pick one of magpie's models for it, give it Library MCP servers and skills, and switch it off to put its config back.", - "zh": "把 nanobot(HKUDS/nanobot)加到 magpie 的 Agent 页:给它选 magpie 的模型、勾选资源库的 MCP 服务与技能,关掉即恢复它原来的配置。" - } - }, - { - "package": "@magpie-community/agent-reasonix-desktop", - "name": "Reasonix 1.x", - "kind": "agent", - "community": true, - "icon": "reasonix-color", - "summary": { - "en": "Reasonix 1.x Desktop and CLI on Magpie's models, preserving existing providers. Restart Reasonix after choosing a model; disconnect before uninstalling. Separate from the built-in Reasonix Studio 2.x adapter.", - "zh": "为 Reasonix 1.x 桌面版和 CLI 选择 Magpie 模型,保留原有供应商;选模型后重启 Reasonix,卸载前先断开。与内置的 Reasonix Studio 2.x 适配器分开。" - } - }, - { - "package": "@magpie-community/agent-deepzora", - "name": "Deepzora", - "kind": "agent", - "community": true, - "icon": "https://raw.githubusercontent.com/SivanCola/deepzora/ad855397cdfdaf2ba7e3e3414583ba777412cbe4/desktop/build/linux/icons/hicolor/128x128/apps/deepzora-desktop.png", - "summary": { - "en": "Deepzora Desktop, CLI and ACP on Magpie's models, with a reversible TOML/dotenv bridge that preserves existing providers. Restart Deepzora after choosing a model; disconnect before uninstalling. Disable the standalone Deepzora plugin before switching to this package.", - "zh": "为 Deepzora 桌面版、CLI 和 ACP 选择 Magpie 模型,通过可恢复的 TOML/dotenv 桥接保留原有供应商。选模型后重启 Deepzora,卸载前先断开;从独立 Deepzora 插件迁移时先停用旧包。" - } - }, - { - "package": "@magpie-community/agent-pig", - "name": "PiG", - "kind": "agent", - "community": true, - "summary": { - "en": "PiG (Pi in Go): add Magpie's endpoint and model catalog, choose models in PiG, and keep startup settings unchanged.", - "zh": "为 PiG(Pi in Go)注入 Magpie 端点与模型列表,在 PiG 内选择模型,保持启动设置不变。" - } - }, - { - "package": "@magpie-community/agent-bub", - "name": "Bub", - "kind": "agent", - "community": true, - "summary": { - "en": "Bub (bub.build) on magpie's Agents page: pick one of magpie's models and Bub uses it, through ~/.bub/config.yml. Needs Bub 0.5.1 or later.", - "zh": "把 Bub(bub.build)加到 magpie 的 Agent 页:给它选一个 magpie 的模型,通过 ~/.bub/config.yml 生效。需要 Bub 0.5.1 或更新版本。" - } - }, - { - "package": "@magpie-community/agent-type4me", - "name": "Type4Me", - "kind": "agent", - "community": true, - "summary": { - "en": "Type4Me (macOS voice input) on magpie's Agents page: pick one of magpie's models for its LLM post-processing, used from the next dictation.", - "zh": "把 Type4Me(macOS 语音输入)加到 magpie 的 Agent 页:给它的 LLM 后处理选 magpie 的模型,下一次语音输入即生效。" - } - }, - { - "package": "@magpie-community/agent-cowork", - "name": "Tuanjie Cowork", - "kind": "agent", - "community": true, - "summary": { - "en": "Tuanjie Cowork (codely) on magpie's Agents page: pick one of magpie's models and Cowork chats with it, on the subscriptions magpie already holds.", - "zh": "把团结 Cowork(codely)加到 magpie 的 Agent 页:给它选一个 magpie 的模型,它聊天就走那个模型,用的是 magpie 里已登录的那些订阅的额度。" - } - }, - { - "package": "@magpie-community/agent-opencodereview", - "name": "Open Code Review", - "kind": "agent", - "community": true, - "summary": { - "en": "Open Code Review (Alibaba's ocr CLI) on magpie's Agents page: pick one of magpie's models and ocr review / ocr scan use it, read from the config on the next run.", - "zh": "把 Open Code Review(阿里巴巴的 ocr CLI)加到 magpie 的 Agent 页:给它选 magpie 的模型,下一次 ocr review / ocr scan 即从配置中读到并生效。" - } - }, - { - "package": "@magpie-community/agent-pi-desktop", - "name": "Pi-Desktop", - "kind": "agent", - "community": true, - "summary": { - "en": "Pi-Desktop on magpie's Agents page: magpie keeps a Pi-Desktop plugin folder that adds it as a model provider, with its current models and their thinking levels. Pi-Desktop picks up model changes when it starts.", - "zh": "把 Pi-Desktop 加到 magpie 的 Agent 页:magpie 维护一个 Pi-Desktop 插件文件夹,把自己加为模型服务,带上当前模型和推理等级;模型有变化时,Pi-Desktop 重启后生效。" - } - }, - { - "package": "@magpie-community/agent-junie", - "name": "Junie CLI", - "kind": "agent", - "community": true, - "icon": "https://junie.jetbrains.com/favicons/favicon.svg", - "summary": { - "en": "An explicit seed copies Magpie's full model catalog into Junie CLI and uses Magpie's selected model as the launch default, without LiteLLM. Native ECO is unavailable for custom profiles in Junie 3893.6.", - "zh": "手动 seed 向 Junie CLI 同步全部模型,并将 Magpie 所选模型设为启动默认模型,无需 LiteLLM。Junie 3893.6 的自定义 profile 不支持原生 ECO。" - } - }, - { - "package": "@magpie-community/middleware-reserve", - "name": "Reserve", - "kind": "middleware", - "community": true, - "summary": { - "en": "Two routings that keep something in reserve: fill accounts up to a ceiling of their allowance, or use subscriptions before API keys.", - "zh": "两种留余量的路由:账号用到额度上限就换下一个,或订阅账号优先于 API key。" - } - }, - { - "package": "@magpie-community/tools-search1api", - "name": "Search1API", - "icon": "https://s1.dev/icon.png", - "kind": "tools", - "community": true, - "summary": { - "en": "Search1API web search, news, agentic search, page retrieval, sitemaps and trends through Magpie's tools MCP server. Sign in with OAuth or an API key.", - "zh": "通过 Magpie 的工具 MCP 服务使用 Search1API 网页搜索、新闻、智能搜索、页面读取、站点链接和趋势;支持 OAuth 或 API key 登录。" - } - }, - { - "package": "@guotree/magpie-websearch", - "name": "magpie-websearch", - "icon": "https://raw.githubusercontent.com/guotree/magpie-websearch/main/assets/icon-128.png", - "kind": "tools", - "community": true, - "summary": { - "en": "Adds 19 search APIs to Magpie: AnySearch, Bocha, Jina, Kagi, Keenable, Linkup, Metaso, Moonshot, Moonshot CN, Ollama, Parallel, Perplexity, Querit, SerpApi, Serper, StepFun, TinyFish, Z.ai, Zhipu. 4 support keyless access; 15 require keys.", - "zh": "为 Magpie 网页搜索增加 19 家搜索源:AnySearch、Bocha、Jina、Kagi、Keenable、Linkup、Metaso、Moonshot、Moonshot CN、Ollama、Parallel、Perplexity、Querit、SerpApi、Serper、StepFun、TinyFish、Z.ai 和智谱。4 家支持免 Key,其余 15 家需要 API Key。" - } - }, - { - "package": "magpie-search-mix", - "name": "magpie-search-mix", - "kind": "tools", - "community": true, - "summary": { - "en": "Mixes two search APIs for the AI: more accurate than one. Try DeepSeek + Tavily or You.com.", - "zh": "混合搜索:同时问两家搜索 API,比只用一家更准。如 DeepSeek 搭配 Tavily 或 You.com。" - } - }, - { - "package": "magpie-comfyui-link", - "name": "ComfyUI", - "kind": "tools", - "community": true, - "summary": { - "en": "A bridge between Magpie agents and locally hosted ComfyUI models, exposing native workflows, model resources and generated images through Magpie's tools MCP endpoint.", - "zh": "连接 Magpie Agent 与本地 ComfyUI 模型的桥接层,通过 Magpie 工具 MCP 入口提供原生工作流、模型资源和生成图片。" - } - }, - { - "package": "@magpie-community/tools-specstream", - "name": "SpecStream", - "kind": "tools", - "community": true, - "summary": { - "en": "SpecStream web search and page extraction for every agent: answers with the source page of each value, and the text of up to 20 URLs a call. Needs a SpecStream API key.", - "zh": "给每个 agent 的 SpecStream 网页搜索和网页提取:每个值都带来源页面,一次最多读取 20 个网址的正文。需要 SpecStream API key。" - } - }, - { - "package": "@magpie-community/tools-calc", - "name": "Calc", - "kind": "tools", - "community": true, - "summary": { - "en": "Exact arithmetic, unit conversion and calendar date math as tools for every agent: no network, no key.", - "zh": "给每个 agent 的精确计算、单位换算和日历日期计算工具:不联网,不需要 key。" - } - }, - { - "package": "@magpie-community/tools-layerfs", - "name": "LayerFS", - "kind": "tools", - "community": true, - "summary": { - "en": "LayerFS workspaces for every agent: deduplicated checkpoints and branches behind stable filesystem paths, instead of a fresh copy of the tree for every agent.", - "zh": "给每个 agent 的 LayerFS 工作区:稳定的文件系统路径背后是可去重的检查点与分支,而不是每个 agent 各复制一棵项目树。" - } - }, - { - "package": "@magpie-community/page-usage", - "name": "Usage Glance", - "kind": "page", - "community": true, - "summary": { - "en": "A page in magpie's window with calls, errors, tokens and cost for today, 7 or 30 days, and the models behind them. Reads only your usage summary.", - "zh": "在 magpie 窗口里一眼看到今天、7 天或 30 天的调用、错误、token、费用和对应模型。只读取用量汇总。" - } - }, - { - "package": "@magpie-community/page-amp-usage", - "name": "Amp Usage", - "kind": "page", - "community": true, - "icon": "https://raw.githubusercontent.com/magpie-community/plugins/main/packages/amp-usage/page/icon.svg", - "summary": { - "en": "An Amp usage board: requests, tokens and cache hit rate by hour, day and model, read from your Amp account and legacy local threads. Its background service syncs that usage into Magpie requests and Day card, with duplicate checks; set sync to false to only view it.", - "zh": "Amp 用量看板:请求数、Token、缓存命中率,按小时、按日和按模型查看,数据来自 Amp 账号和旧版本地线程。后台服务把这些用量同步进 Magpie 请求统计和每日卡片,带去重检查;把 sync 设为 false 则只查看。" - } - }, - { - "package": "@magpie-community/page-day-card", - "name": "Day Card", - "kind": "page", - "community": true, - "summary": { - "en": "A day of your AI usage printed as a receipt: its cost, requests and streak, the day against your average day, each agent and its models, and its tokens. Copy it as a picture or save it as a PNG. Reads only your usage by day.", - "zh": "把一天的 AI 用量打印成一张小票:费用、请求数、连续使用天数、与日均的倍数、每个 Agent 和它用的模型,以及 token 用量。可复制为图片或保存为 PNG。只读取按天的用量。" - } - }, - { - "package": "@magpie-community/page-fingerprint", - "name": "Fingerprint", - "kind": "page", - "community": true, - "summary": { - "en": "Check for suspected model substitution by comparing response fingerprints across Magpie providers. Uses lm-detector with shared challenges, local evidence history and JSON export; sampling uses provider quota. Reference data is downloaded on demand from lm.ikale.io, falling back to cdn.jsdmirror.com.", - "zh": "在 Magpie 内检测响应指纹,为疑似模型替换提供核查线索。基于 lm-detector,支持跨渠道共享挑战、保留本地证据与 JSON 导出;采样会消耗模型额度。参考库按需从 lm.ikale.io 下载,备用源为 cdn.jsdmirror.com。" - } - }, - { - "package": "@magpie-community/page-feishu-sidebar", - "name": "Feishu sidebar", - "kind": "page", - "community": true, - "summary": { - "en": "Generate an AppLink to open Magpie Web in the Feishu desktop sidebar. The page uses the Web address you enter and reads no Magpie data.", - "zh": "生成飞书链接,在飞书桌面端左侧导航打开 Magpie Web。页面使用你填写的 Web 地址,不读取 Magpie 数据。" - } - }, - { - "package": "@magpie-community/page-chat", - "name": "Chat", - "kind": "page", - "community": true, - "summary": { - "en": "A chat page in magpie's window by Magic_zero: talk to magpie's own models with your MCP servers, instructions and skills, images and LaTeX. Runs a local service on 127.0.0.1 that only its own page can use.", - "zh": "Magic_zero 做的 magpie 窗口内「对话」页面:用 magpie 自己的模型聊天,带上你的 MCP 服务器、指令和技能,支持图片与 LaTeX。本地服务只在 127.0.0.1,只给自己的页面用。" - } - }, - { - "package": "@magpie-community/page-agent-office", - "name": "Agent Office", - "kind": "page", - "community": true, - "summary": { - "en": "Reads the session files of Claude Code, Codex, Copilot CLI, Kimi Code, DeepSeek Harness, Gemini CLI, Qwen Code, Pi, omp, Qoder, WorkBuddy, CodeBuddy and Grok Build on this computer (prompts, replies and tool commands; text can be turned off). Shows them in a real WebGL 3D office with eleven scenes, an isometric/2D office, or a grouped workbench with Today usage, recent conversations and approvals.", - "zh": "读取本机 Claude Code、Codex、Copilot CLI、Kimi Code、DeepSeek Harness、Gemini CLI、Qwen Code、Pi、omp、Qoder、WorkBuddy、CodeBuddy、Grok Build 的会话文件(含提问、回复与工具命令,可关闭文本),用真 3D 办公室(十一种场景)、等距/2D 办公室或分组工作台展示,带今日用量、最近对话与批准状态。" - } - }, - { - "package": "@magpie-community/page-usage-tree", - "name": "Account usage", - "kind": "page", - "community": true, - "summary": { - "en": "A provider -> account -> model usage tree with per-account calls, errors and cost. Adds the account layer that Usage Glance's usage:summary does not contain, read from Magpie's local usage API.", - "zh": "按供应商 -> 账号 -> 模型展示用量,包含每个账号的调用、错误和费用;补充 Usage Glance 的 usage:summary 所没有的账号层,数据来自 Magpie 本地用量 API。" - } - }, - { - "package": "@magpie-community/agent-kimi-desktop", - "name": "Kimi", - "kind": "agent", - "community": true, - "summary": { - "en": "Kimi desktop Work with Magpie models: configure primary and secondary models on its embedded page, select native subagent roles, and disconnect to restore the original service. Not Kimi Code CLI.", - "zh": "普通 Kimi 桌面版 Work 使用 Magpie 模型:在内嵌页面配置主、副模型,为原生子代理选择角色,断开时恢复原服务。不是 Kimi Code CLI。" - } - }, - { - "package": "@magpie-community/agent-minimax-design", - "name": "MiniMax Design", - "kind": "agent", - "community": true, - "icon": "https://design.minimax.io/favicon.ico", - "summary": { - "en": "MiniMax Design (the local multimodal creation studio) on magpie's Agents page: one switch points Design's custom text model at magpie's gateway, and Design's model menu follows magpie's model list automatically. Quit Design before switching; Windows only (see COMPATIBILITY.md).", - "zh": "把 MiniMax Design(本地多模态创作工作室)加到 magpie 的 Agent 页:一个开关把 Design 的自定义文本模型指到 magpie 网关,Design 的模型菜单随 magpie 的模型列表自动同步。切换前先退出 Design;仅支持 Windows,详见 COMPATIBILITY.md。" - } - }, - { - "package": "@magpie-community/gateway-notify", - "name": "Notify", - "kind": "middleware", - "community": true, - "summary": { - "en": "A push when an agent's turn is done or fails, after it ran long enough to be worth one: ntfy, Bark, Gotify, Server酱, Feishu, DingTalk, WeCom, Telegram or your own webhook. Also pushes when a subscription account is signed out and needs signing in again. Agent, project, model and time only by default; the reply's text is opt-in.", - "zh": "agent 一轮任务跑完或失败时推送提醒(只提醒跑够设定时长的任务):支持 ntfy、Bark、Gotify、Server酱、飞书、钉钉、企业微信、Telegram 和自定义 webhook。订阅账号掉线需要重新登录时也会推送。默认只发 agent、项目、模型和用时,回复正文需手动开启。" - } - }, - { - "package": "@magpie-community/hooks-mail-watch", - "name": "Mail Watch", - "kind": "middleware", - "community": true, - "summary": { - "en": "Email alerts when a provider call fails or a quota window runs out: your own SMTP account, your own words in the mail, and your own list of what to watch, with cooldowns and an hourly cap so a bad hour stays bounded — plus an optional periodic usage digest.", - "zh": "供应商调用失败或额度窗口用尽时发邮件提醒:用你自己的 SMTP 账号、自己的邮件文案、自己勾选的监控事项;同类事件冷却合并,另有每小时上限,糟糕的一小时不会失控;另可开启定期用量摘要。" - } - }, - { - "package": "@magpie-community/opencode-cloudflare-workers-ai-auth", - "name": "Cloudflare Workers AI", - "providers": [ - "cloudflare-workers-ai" - ], - "community": true, - "summary": { - "en": "Cloudflare Workers AI with your API token (Workers AI:Edit): 28+ models discovered live from models.dev, the 8 paid-only models flagged, and a Neurons/day allowance row in magpie quota.", - "zh": "用你的 Cloudflare API Token(Workers AI:Edit 权限)登录 Workers AI:28+ 模型经 models.dev 实时发现,付费专属模型自动标注,magpie quota 显示每日 Neurons 免费额度。" - } - }, - { - "package": "@magpie-community/opencode-openai-decisions-auth", - "name": "OpenAI Decisions", - "providers": [ - "openai-decisions" - ], - "community": true, - "icon": "data:image/svg+xml,%3Csvg%20xmlns%3D%22http%3A%2F%2Fwww.w3.org%2F2000%2Fsvg%22%20viewBox%3D%220%200%2064%2064%22%3E%3Crect%20width%3D%2264%22%20height%3D%2264%22%20rx%3D%2214%22%20fill%3D%22%23111827%22%2F%3E%3Cpath%20d%3D%22M20%2032h12V20h12M32%2032v12h12%22%20fill%3D%22none%22%20stroke%3D%22white%22%20stroke-width%3D%225%22%20stroke-linecap%3D%22round%22%20stroke-linejoin%3D%22round%22%2F%3E%3Ccircle%20cx%3D%2220%22%20cy%3D%2232%22%20r%3D%225%22%20fill%3D%22white%22%2F%3E%3Ccircle%20cx%3D%2246%22%20cy%3D%2220%22%20r%3D%225%22%20fill%3D%22white%22%2F%3E%3Ccircle%20cx%3D%2246%22%20cy%3D%2244%22%20r%3D%225%22%20fill%3D%22white%22%2F%3E%3C%2Fsvg%3E", - "summary": { - "en": "OpenAI's native Decisions API as System One for Magpie classifiers and decision groups: choice, score and boolean probabilities, with standard API-key sign-in and per-question refusal handling.", - "zh": "将 OpenAI 原生 Decisions API 转为 System One,供 Magpie 分类器与决策模型组使用,保留选择、评分和布尔概率;使用标准 API key 登录,单题拒绝时保留其余答案。" - } - }, - { - "package": "@magpie-community/opencode-openrouter-decisions-auth", - "name": "OpenRouter Decisions", - "providers": [ - "openrouter-decisions" - ], - "community": true, - "icon": "data:image/svg+xml,%3Csvg%20xmlns%3D%22http%3A%2F%2Fwww.w3.org%2F2000%2Fsvg%22%20viewBox%3D%220%200%2064%2064%22%3E%3Ccircle%20cx%3D%2232%22%20cy%3D%2232%22%20r%3D%2229%22%20fill%3D%22%230f172a%22%2F%3E%3Cpath%20d%3D%22M20%2040a12%2012%200%201%201%2012%2012%22%20fill%3D%22none%22%20stroke%3D%22%2365a3ff%22%20stroke-width%3D%225%22%20stroke-linecap%3D%22round%22%2F%3E%3Ccircle%20cx%3D%2244%22%20cy%3D%2228%22%20r%3D%226%22%20fill%3D%22%2365a3ff%22%2F%3E%3C%2Fsvg%3E", - "summary": { - "en": "OpenRouter's Decisions API as System One for Magpie classifiers and decision groups: Microsoft Decision 1.1 and 1, with choice, score and boolean probabilities, credit usage and per-question refusal handling.", - "zh": "将 OpenRouter Decisions API 转为 System One,供 Magpie 分类器与决策模型组使用:Microsoft Decision 1.1 与 1,保留选择、评分和布尔概率、额度用量,单题拒绝时保留其余答案。" - } - }, - { - "package": "@magpie-community/opencode-river-auth", - "name": "River", - "providers": [ - "river" - ], - "community": true, - "summary": { - "en": "River (river.ai) pay-as-you-go inference with your API key: per-account models from /v1/models, documented reasoning efforts and live prices from River's public pricing endpoint.", - "zh": "用你的 API key 接入 River (river.ai) 按量计费推理:按账号从 /v1/models 列出模型,带官方推理档位与实时价格。" - } - }, - { - "package": "@magpie-community/opencode-sap-aicore-auth", - "name": "SAP AI Core", - "providers": [ - "sap-aicore" - ], - "community": true, - "summary": { - "en": "Existing SAP AI Core deployments with client-secret credential-file sign-in, per-resource-group GPT Responses and Claude Messages discovery, streaming and agent tools. No deployment provisioning or quota reporting; Magpie only.", - "zh": "用本机 client-secret 凭据文件连接已有 SAP AI Core 部署:按资源组发现 GPT Responses 与 Claude Messages 模型,支持流式回复和 agent 工具。不创建部署,不报告额度;仅验证 Magpie。" - } - }, - { - "package": "magpie-plugin-pink", - "name": "PinK", - "kind": "agent", + { "package": "@magpie-community/opencode-gumloop-auth", "name": "Gumloop", "icon": "https://www.gumloop.com/favicon.ico", "providers": ["gumloop"], "community": true, + "summary": { "en": "Gumloop with your API key and user ID: account models, streaming, tool calls and a locally calibrated credit estimate.", "zh": "使用 API key 和用户 ID 接入 Gumloop:按账号发现模型,支持流式响应、工具调用和本地校准的积分估算。" } }, + { "package": "@magpie-community/opencode-agnes-auth", "name": "Agnes AI", "icon": "https://www.agnes-ai.com/images/biglogo.png", "providers": ["agnes"], "community": true, + "summary": { "en": "Agnes AI international with Chrome browser sign-in: live OpenAI-compatible text models, Token Plan and Agent Plan usage, and the USD wallet balance.", "zh": "通过 Chrome 浏览器登录 Agnes AI 国际版:动态发现 OpenAI 兼容文本模型,显示 Token Plan、Agent Plan 用量与美元余额,无需手动复制会话 token。" } }, + { "package": "@magpie-community/opencode-atomcode-auth", "name": "AtomCode", "icon": "https://atomgit.com/favicon.ico", "providers": ["atomcode"], "community": true, + "summary": { "en": "Unofficial AtomGit CodingPlan access for magpie and OpenCode: browser OAuth or local AtomCode CLI sign-in, signed OpenAI-compatible requests, live models, quota windows, and a read-only diagnostics page.", "zh": "非官方接入 AtomGit CodingPlan:支持浏览器 OAuth 或复用本机 AtomCode CLI 登录,发送签名请求,显示实时模型与额度窗口,并提供只读诊断页。" } }, + { "package": "@magpie-community/gateway-powercontext", "name": "PowerContext", "kind": "middleware", "community": true, + "icon": "https://raw.githubusercontent.com/oceanbase/powercontext/master/src/powercontext/server/dashboard/static/brand/favicon-color.svg", + "summary": { "en": "Recall from your PowerContext service and optionally save complete conversations. Configure the service URL first; capture is off by default.", "zh": "从自己的 PowerContext 服务召回记忆,并按需存回完整对话。先配置服务地址;对话存回默认关闭。" } }, + { "package": "@magpie-community/gateway-nowledge-mem", "name": "Nowledge Mem", "kind": "middleware", "community": true, + "summary": { "en": "Give every harness behind Magpie Nowledge Mem automatically: recall and the Context Bundle (or Working Memory) before each request, per-harness provenance, optional import of completed conversations as Mem Threads, and a real connection test in the page. A local or remote Mem; saving conversations is off by default.", "zh": "让 Magpie 后面的每个 harness 自动接入 Nowledge Mem:每次请求前召回记忆与 Context Bundle(缺失时用 Working Memory),记录来自哪个 harness,按需把已完成的对话导入为 Mem Thread,并在页面中提供真实的连接测试。本机或远端 Mem 都可用;对话导入默认关闭。" } }, + { "package": "@magpie-community/gateway-aelios", "name": "Aelios", "kind": "middleware", "community": true, + "summary": { "en": "Long-term memory from your own Aelios on Cloudflare: recall before each turn for every agent behind magpie, and optionally record the conversation so Aelios can learn it. Set it up on the plugin's own page, with a connection test and per-model or per-agent routing to separate assistants. Recording is off by default.", "zh": "接入你自己部署在 Cloudflare 上的 Aelios 长期记忆:magpie 后面的每个 agent 每轮自动召回,按需把对话交给 Aelios 记住。在插件自己的页面里配置,可选择连接测试,并按模型或客户端分流到不同的助手。记录默认关闭。" } }, + { "package": "@magpie-community/gateway-respire-memory", "name": "respire-memory", "kind": "middleware", "community": true, + "icon": "https://raw.githubusercontent.com/risense-ai/respire-magpie/v0.3.0/page/icon.svg", + "summary": { "en": "Help AI agents evolve through experience: per-turn recall and complete engine-policy reads, automatic process diaries and grounded archival with write readback. Configure an existing local Runtime and generation model first. Archival defaults on; raw capture off. Background storage happens after reply completion.", "zh": "让 AI Agent 通过经验积累持续自进化:每轮召回并读完整政策,自动保存过程日记与整理长期记忆,写后回读。先配置已有本地 Runtime 和生成模型;归档默认开启,原始采集关闭。宿主限制下存储在回复完成后执行。" } }, + { "package": "@openviking/magpie-plugin", "name": "OpenViking", "kind": "middleware", "community": true, + "icon": "https://raw.githubusercontent.com/volcengine/OpenViking/main/docs/images/ov-logo-icon.png", + "summary": { "en": "Recall OpenViking context before model requests and optionally capture completed conversations. Reuses your existing OpenViking configuration; capture is off by default. Maintained by the OpenViking project.", "zh": "在模型请求前召回 OpenViking 上下文,并按需采集完整对话。复用现有 OpenViking 配置;采集默认关闭。由 OpenViking 项目维护。" } }, + { "package": "@vshulcz/magpie-deja", "name": "deja-vu", "kind": "middleware", "community": true, + "icon": "https://raw.githubusercontent.com/vshulcz/deja-vu/main/docs/assets/icon.svg", + "summary": { "en": "Recall local session history from the transcripts Claude Code, Codex, Hermes, Gemini CLI, OpenCode, Cursor and DSH write on this machine before each request. Scopes lookup to the request's project directory (ctx.cwd) by default and scrubs secret patterns before attaching recalled excerpts to the request sent to your model provider. When recalled history helped, the model is asked to end its reply with a one-line \"déjà vu: … — reusing it.\" note.", "zh": "每次请求前从本机 Claude Code、Codex、Hermes、Gemini CLI、OpenCode、Cursor 与 DSH 的磁盘会话记录中召回相关历史。默认严格限定在请求所属的项目目录(ctx.cwd)内检索,并对密钥与令牌执行脱敏后,将召回片段随请求一并发送给所配置的模型供应商。当召回的历史记录有帮助时,会请模型在回复末尾附上一行“déjà vu: … — reusing it.”说明。" } }, + { "package": "@magpie-community/opencode-copilot-auth", "name": "GitHub Copilot", "icon": "githubcopilot", "providers": ["github-copilot"], "community": true, + "summary": { "en": "GitHub Copilot on GitHub.com or GitHub Enterprise, with premium request or credit usage, remaining allowance and quota reset dates.", "zh": "GitHub Copilot:支持 GitHub.com 与 GitHub Enterprise,显示高级请求或积分用量、剩余额度和重置时间。" } }, + { "package": "@magpie-community/opencode-joycode-auth", "name": "JoyCode", "providers": ["joycode"], "community": true, + "summary": { "en": "Unofficial JoyCode provider: JD account browser sign-in, live models and credit multipliers, plan credits and streaming tool calls.", "zh": "非官方 JoyCode 插件:京东账号浏览器登录,动态模型与积分倍率、套餐额度及流式工具调用。" } }, + {"package": "@magpie-community/opencode-doubao-work-auth", "name": "豆包工作", "icon": "doubao-color", "providers": ["doubao-work"], "community": true, "summary": {"en": "Not a model API: Doubao Work has none for other clients. This drives its desktop app (or Doubao's) the way a person does, so each request is one run of Doubao's cloud agent in a new task, with the client's tools asked for as text. macOS and Windows; unofficial.", "zh": "不是模型接口:豆包工作没有给第三方客户端用的模型接口。本插件像人一样操作豆包工作(或豆包)桌面版,每个请求是豆包云端智能体在一个新任务里的一次执行,客户端的工具以文本方式调用。macOS 和 Windows;非官方。"}}, + {"package": "@magpie-community/opencode-kimi-work-auth", "name": "Kimi Work", "icon": "kimi", "providers": ["kimi-work"], "community": true, "summary": {"en": "Kimi desktop Work subscription through its local sign-in, with live Work models and shared membership usage. Separate from Kimi Code.", "zh": "使用本机普通 Kimi 桌面版的 Work 登录,自动读取 Work 模型和会员共享用量;独立于 Kimi Code。"}}, + { "package": "@magpie-community/opencode-zen-free-auth", "name": "OpenCode Zen Free", "icon": "opencode", "providers": ["opencode-zen-free"], "community": true, + "summary": { "en": "OpenCode Zen's free models with the public credential, automatic discovery and native streaming protocols.", "zh": "使用 public 凭证接入 OpenCode Zen 免费模型,自动发现模型并支持原生流式协议。" } }, + { "package": "@magpie-community/opencode-zcode-auth", "name": "ZCode", "icon": "zcode", "providers": ["zcode"], "community": true, "replaces": "zcode", + "summary": { "en": "Z.ai and BigModel GLM Coding Plan, team seats and ZCode's free Start Plan, with a daily claim of its gift plans.", "zh": "Z.ai 与智谱 GLM Coding Plan,支持团队席位和 ZCode 免费 Start Plan,可每日领取 ZCode 活动包。" } }, + { "package": "@magpie-community/opencode-qoder-auth", "name": "Qoder", "icon": "qoder", "providers": ["qoder", "qoder-cn", "qwenwork", "qwenwork-enterprise"], "community": true, "replaces": "qoder", + "summary": { "en": "Your Qoder subscription: international (qoder.com) or Qoder CN, the China site (qoder.cn, Alibaba Cloud or phone sign-in), each its own provider. Also QwenWork (千问办公): the public cloud or the enterprise flagship with its own domain. Unofficial for QwenWork.", "zh": "用 Qoder 自己的设备登录使用 Qoder 订阅:国际版(qoder.com)与 Qoder 中国版(Qoder CN,qoder.cn,阿里云账号 / 手机号登录)各是一个订阅;也接入千问办公:公有云与企业旗舰版(输入企业自己的域名)。千问办公部分非官方。" } }, + { "package": "@magpie-community/opencode-devin-auth", "name": "Devin", "icon": "devin", "providers": ["devin"], "community": true, "replaces": "devin", + "summary": { "en": "Devin's models with your Devin account — a browser sign-in or the devin CLI's.", "zh": "用 Devin 账号使用 Devin 的模型:浏览器登录,或沿用 devin CLI 的登录。" } }, + { "package": "@magpie-community/opencode-zed-auth", "name": "Zed", "icon": "zed", "providers": ["zed"], "community": true, "replaces": "zed", + "summary": { "en": "Claude, GPT, Gemini and Grok hosted by Zed, on your Zed plan.", "zh": "用 Zed 订阅使用 Zed 托管的 Claude、GPT、Gemini 和 Grok。" } }, + { "package": "@magpie-community/opencode-factory-auth", "name": "Factory", "icon": "factory", "providers": ["factory"], "community": true, "replaces": "factory", + "summary": { "en": "Factory's Droid subscription: models from Anthropic, OpenAI, xAI and open labs.", "zh": "Factory Droid 订阅:Anthropic、OpenAI、xAI 及开源模型。" } }, + { "package": "@magpie-community/opencode-grok-auth", "name": "Grok", "icon": "xai", "providers": ["grok"], "community": true, "replaces": "grok", + "summary": { "en": "SuperGrok and X Premium+, through the Grok Build CLI's sign-in.", "zh": "通过 Grok Build CLI 的登录使用 SuperGrok 和 X Premium+。" } }, + { "package": "@magpie-community/opencode-commandcode-auth", "name": "Command Code", "icon": "commandcode", "providers": ["commandcode-plan"], "community": true, "replaces": "commandcode-plan", + "summary": { "en": "Command Code plans — Pro, GOAT, Max, Ultra, Go and Teams.", "zh": "Command Code 各档套餐:Pro、GOAT、Max、Ultra、Go 和 Teams。" } }, + { "package": "@magpie-community/opencode-mimo-auth", "name": "Xiaomi MiMo", "icon": "mimocode", "providers": ["mimo-app"], "community": true, "replaces": "mimo-app", + "summary": { "en": "Xiaomi MiMo models with your Xiaomi account.", "zh": "用小米账号使用小米 MiMo 模型。" } }, + { "package": "@magpie-community/opencode-cursor-auth", "name": "Cursor", "icon": "cursor", "providers": ["cursor"], "community": true, "replaces": "cursor", + "summary": { "en": "Cursor Pro, Pro+, Ultra and Teams — every model your plan has, with tool calls.", "zh": "Cursor Pro、Pro+、Ultra 与 Teams:套餐里的所有模型,支持工具调用。" } }, + { "package": "@magpie-community/opencode-kiro-auth", "name": "Kiro", "icon": "kiro-color", "providers": ["kiro"], "community": true, "replaces": "kiro", + "summary": { "en": "Kiro Free, Pro, Pro+ and Power — Kiro's sign-in, kiro-cli's, an API key, or accounts imported by refresh token.", "zh": "Kiro Free、Pro、Pro+ 与 Power:Kiro 登录、沿用 kiro-cli 的登录、API key,或按 refresh token 导入账号。" } }, + { "package": "@magpie-community/opencode-minimax-auth", "name": "MiniMax Code", "icon": "minimax-color", "providers": ["minimax-code", "minimax-code-global"], "community": true, + "summary": { "en": "MiniMax Code credits and M Plan with device sign-in, native web search, image/video generation and media tools through Magpie MCP.", "zh": "MiniMax Code 积分与 M Plan、设备登录,提供原生网络搜索、生图、生视频及多模态 MCP 工具。" } }, + { "package": "@magpie-community/opencode-warp-auth", "name": "Warp", "icon": "warp", "providers": ["warp"], "community": true, + "summary": { "en": "Warp (the terminal) AI request credits through the Warp app's sign-in: the auto routers and every model family the account lists.", "zh": "用 Warp 终端应用的登录使用 Warp AI 请求额度:auto 路由与账户可用的全部模型家族。" } }, + { "package": "@magpie-community/opencode-workbuddy-auth", "name": "WorkBuddy", "icon": "workbuddy-color", "providers": ["workbuddy", "workbuddy-ai"], "community": true, "replaces": "workbuddy", + "summary": { "en": "WorkBuddy's China build (CodeBuddy plan) and its international build.", "zh": "WorkBuddy 国内版(CodeBuddy 套餐)与国际版。" } }, + { "package": "@magpie-community/opencode-cline-auth", "name": "Cline", "icon": "cline", "providers": ["cline"], "community": true, + "summary": { "en": "Cline's models on your own Cline account — usage billing, ClinePass and Cline's free models — with Cline's device sign-in or an API key.", "zh": "用自己的 Cline 账号使用 Cline 的模型——用量计费、ClinePass 与免费模型——支持 Cline 设备登录或 API key。" } }, + { "package": "@magpie-community/opencode-trae-auth", "name": "Trae", "icon": "trae", "providers": ["trae-cn", "trae-global"], "community": true, + "summary": { "en": "Trae CN (trae.cn) and Trae international (trae.ai) models with your own account, free tier included — Trae's own browser sign-in. Experimental.", "zh": "用 Trae CN(trae.cn)与 Trae 国际版(trae.ai)账号使用各自的模型,含免费版,使用 Trae 自己的浏览器登录。实验性。" } }, + { "package": "@magpie-community/opencode-opendesign-auth", "name": "Open Design", "providers": ["opendesign"], "community": true, + "summary": { "en": "Open Design Cloud's Design Plan (open-design.ai): the account's device sign-in, the models the plan entitles it to, and the plan's 5-hour and 7-day allowance and wallet balance.", "zh": "Open Design Cloud 的 Design Plan(open-design.ai):账号设备登录、套餐可用的模型,以及 5 小时与 7 天额度、钱包余额。" } }, + { "package": "@magpie-community/opencode-catpaw-auth", "name": "CatPaw", "providers": ["catpaw"], "community": true, + "summary": { "en": "Meituan's 妙手 CatPaw (ai.catpaw.meituan.com): a browser web sign-in or the desktop client's saved credentials, chat through its round/event/turn conversation protocol, models and credit balance.", + "zh": "美团妙手 CatPaw(ai.catpaw.meituan.com):浏览器网页登录或妙手桌面端已保存的登录态,经其 round/event/turn 会话协议对话,含模型与积分余额。" } }, + { "package": "@magpie-community/opencode-loomy-auth", "name": "Loomy", "providers": ["loomy"], "community": true, + "summary": { "en": "iFlytek's Loomy (loomyad.xunfei.cn): SMS sign-in, OpenAI-compatible chat with reasoning-effort folding and image markers, the live model catalog, two point wallets and the daily first-login grant.", + "zh": "讯飞 Loomy(loomyad.xunfei.cn):短信验证码登录,OpenAI 兼容对话(思考档折叠与图片标记适配),实时模型目录、双积分钱包与每日首登赠送。" } }, + { "package": "@magpie-community/opencode-nous-auth", "name": "Nous Portal (OAuth)", "icon": "nousresearch", "providers": ["nous-portal"], "community": true, + "summary": { "en": "Nous Portal models, with Nous Portal's device sign-in.", "zh": "通过 Nous Portal 自己的设备码登录使用 Nous Portal 模型。" } }, + { "package": "@magpie-community/opencode-raccoon-auth", "name": "商汤小浣熊", "icon": "sensenova-color", "providers": ["raccoon"], "community": true, + "summary": { "en": "SenseTime Raccoon (xiaohuanxiong.com) models, account credits and daily check-in, with browser sign-in or Raccoon desktop's live sign-in.", "zh": "商汤小浣熊(xiaohuanxiong.com)模型、积分余额与每日签到,支持浏览器登录或直接复用小浣熊桌面端登录态。" } }, + { "package": "@magpie-community/opencode-sensenova-auth", "name": "SenseNova", "icon": "sensenova-color", "providers": ["sensenova-token-plan"], "community": true, + "summary": { "en": "SenseNova (SenseTime 日日新) Token Plan: a browser sign-in (or the console account and password) reads the credit pools, their 5-hour and 7-day windows and the expiring grants, and serves the plan's own models — the browser sign-in reuses or creates a Token Plan API key for you.", "zh": "商汤日日新 Token Plan:浏览器登录(或控制台账号密码)读取积分池、5 小时 / 7 天窗口与将过期的返赠余额,并直接调用计划内模型——浏览器登录会自动复用或创建 Token Plan API Key。" } }, + { "package": "@magpie-community/opencode-inkstone-auth", "name": "书生·端砚", "icon": "https://raw.githubusercontent.com/magpie-community/plugins/main/packages/inkstone/icon.svg", "providers": ["intern-inkstone"], "community": true, + "summary": { "en": "Unofficial: Intern InkStone(书生·端砚) Token Plan: a browser sign-in reads the ink-credit balance, the 5-hour and 7-day windows and the credit packs, and creates an ink-* API key (or reuses a pasted one) to serve the plan's models for chat.", "zh": "非官方:书生·端砚(Intern InkStone)墨点计划:浏览器登录读取墨点余额、5 小时 / 7 天窗口与墨点包,并创建 ink-* API Key(或复用现有 Key)直接调用计划内模型对话。" } }, + { "package": "@magpie-community/opencode-autoclaw-auth", "name": "AutoClaw", "providers": ["autoclaw", "autoclaw-cn"], "community": true, + "summary": { "en": "AutoClaw (Zhipu OpenClaw) cloud accounts: oversea web sign-in (Google/z.ai) or CN phone SMS, credit balance and daily check-in.", "zh": "AutoClaw(智谱澳龙/OpenClaw)云端账号:国际版网页登录(Google/z.ai)或国内手机验证码,积分余额与每日签到。" } }, + { "package": "@magpie-community/opencode-codearts-auth", "name": "CodeArts / OfficeAce", "providers": ["codearts", "officeace"], "community": true, + "summary": { "en": "Unofficial: Huawei Cloud CodeArts (码道) coding plans and OfficeAce (AgentArts), each signed in with Huawei Cloud OAuth on its own card — GLM, DeepSeek, openPangu and OfficeAce's MaaS models, live models and credit rates, credit usage by bucket and the daily check-in. Replaces officeace-auth; accounts signed in there keep working.", "zh": "非官方:华为云码道(CodeArts)编码套餐与 OfficeAce(AgentArts),两张卡各自用华为云 OAuth 登录;GLM、DeepSeek、openPangu 与 OfficeAce 的 MaaS 模型,实时模型与积分倍率、分桶积分用量和每日签到。取代 officeace-auth,原已登录的账号继续可用。" } }, + { "package": "@magpie-community/opencode-gcp-ai-auth", "name": "GCP AI", "providers": ["gcp-ai"], "community": true, + "summary": { "en": "Vertex AI models and Google Cloud promotional credits, signed with your own gcloud or ADC login. Lists only the models your project can call.", "zh": "用本机 gcloud / ADC 登录使用 Vertex AI 模型,并读取 Google Cloud 促销赠金;只列出项目实际能调用的模型。" } }, + { "package": "magpie-review-rules", "name": "Claude Desktop Review & Model Rules", "kind": "middleware", + "summary": {"en":"Pin Claude Desktop auto-mode action reviews to a chosen Magpie model, independently of the conversation model. Configure per-model Claude compatibility presets and sandboxed JavaScript request rules in a native settings page. No request changes until configured.","zh":"固定 Claude Desktop 自动模式的操作审核模型,不随对话模型切换;通过原生设置页配置逐模型 Claude 兼容预设与沙箱 JavaScript 请求规则。首次安装不改写请求。"} }, + { "package": "@magpie-community/middleware-param-override", "name": "Param Override", "kind": "middleware", "community": true, + "summary": { "en": "New API's parameter override: set, delete, move or rewrite fields of every request, by model or by condition.", "zh": "New API 的参数覆盖:按模型或条件设置、删除、移动或改写每个请求的字段。" } }, + { "package": "@magpie-community/middleware-model-map", "name": "Model Map", "kind": "middleware", "community": true, + "summary": { "en": "New API's model redirection: send a model under another name, by name or pattern; replies keep the name asked for.", "zh": "New API 的模型重定向:按名称或正则把模型换成另一个发出,回复里仍是原来的名字。" } }, + { "package": "@magpie-community/middleware-system-prompt", "name": "System Prompt", "kind": "middleware", "community": true, + "summary": { "en": "Add your own system prompt to every request, or only some agents' or models', in each API's own place.", "zh": "给每个请求(或指定 agent、模型)加上自己的系统提示词,放在各 API 自己的位置。" } }, + { "package": "@magpie-community/middleware-output-strategy", "name": "Output Strategy", "kind": "middleware", "community": true, + "summary": { "en": "Seven composable output styles, all off by default, with a selection panel and prompt-cache-stable injection that preserves Codex instructions. Unofficial; by manjughosa.", "zh": "七种可组合的输出样式,全部默认关闭;带选择面板,注入不破坏提示缓存,也不改写 Codex instructions。由 manjughosa 独立开发并维护,非官方出品。" } }, + { "package": "@magpie-community/middleware-word-guard", "name": "Word Guard", "kind": "middleware", "community": true, + "summary": { "en": "New API's sensitive-word filter: turn away or mask words and patterns in what you send, and in replies.", "zh": "New API 的敏感词过滤:拒绝或替换你发出的内容和回复里的词与正则。" } }, + { "package": "@magpie-community/middleware-company-guard", "name": "Company Guard", "kind": "middleware", "community": true, + "summary": { "en": "Recognize and mask company names with stable session aliases, restore replies, and show local receipts and explicit Codex session approvals. Unlike Word Guard, names can be restored. Text only; rule-based detection.", "zh": "识别公司名称并使用会话内稳定代号,回复返回本机时还原,提供本地回执与需明确确认的 Codex 会话授权。区别于 Word Guard 的词表拒绝或替换,支持名称还原;仅处理文字,采用规则识别。" } }, + { "package": "@magpie-community/middleware-think-tags", "name": "Think Tags", "kind": "middleware", "community": true, + "summary": { "en": "Take a model's leading , or block out of its reply's text, or put a Chat reply's reasoning into it.", "zh": "去掉回复正文开头的 、 或 思考块,或把 Chat 回复的思考内容放进正文。" } }, + { "package": "@magpie-community/middleware-empty-reply", "name": "Empty Reply", "kind": "middleware", "community": true, + "summary": { "en": "Drop the end of a reply that has reasoning but no text, so Claude Code and Codex retry instead of treating an empty turn as finished. Anthropic and Responses, streamed or whole.", "zh": "丢掉「只有思考、没有正文」回复的结尾,让 Claude Code 和 Codex 重试,而不是把空回合当成正常结束。支持 Anthropic 与 Responses 的流式与整包回复。" } }, + { "package": "@magpie-community/agent-nowledge-mem", "name": "Nowledge Mem / AI Now models", "kind": "agent", "community": true, "icon": "https://mem.nowledge.co/favicon.png", + "summary": {"en": "Choose Nowledge Mem Default, AI Now and Background models through Magpie. Distinct from gateway-nowledge-mem recall/capture. Disconnect before uninstalling.", "zh": "在 Magpie 管理 Nowledge 的 Default、AI Now 和后台模型;与 gateway-nowledge-mem 的记忆召回、对话导入用途不同。卸载前先断开。"} }, + { "package": "@magpie-community/agent-kilo", "name": "Kilo Code", "kind": "agent", "community": true, "icon": "https://github.com/Kilo-Org.png", + "summary": { "en": "Kilo Code (kilo.ai) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put its config back.", "zh": "把 Kilo Code(kilo.ai)加到 magpie 的 Agent 页:选择 magpie 的模型,关掉即恢复原来的配置。" } }, + { "package": "@magpie-community/middleware-antigravity-cloak", "name": "Antigravity Cloak", "kind": "middleware", "community": true, + "summary": { "en": "Claude Code and Claude Desktop on an Antigravity account: the words Antigravity refuses their system prompt for are broken up, in what Antigravity alone is sent.", "zh": "让 Claude Code 和 Claude Desktop 能用 Antigravity 账号:只在发给 Antigravity 的请求里拆开它据以拒绝系统提示词的那几个词。" } }, + { "package": "@magpie-community/middleware-fovea", "name": "Fovea", "kind": "middleware", "community": true, + "summary": { "en": "Thin an agent's old history while preserving a recent window and tool pairing. Optional saccade collapses old content on a new topic; a structured state block replaces the raw task message; cut boundaries move in blocks to keep provider prompt caches settling.", "zh": "裁薄 Agent 旧历史,保留最近窗口与工具配对;可选 saccade 在新话题时折叠旧内容,结构化状态块替代原始任务消息,裁剪边界分块移动以保护 prompt cache。" } }, + { "package": "@magpie-community/middleware-codex-history", "name": "Codex History Filter", "kind": "middleware", "community": true, + "summary": { "en": "Opt-in, per-provider Codex Responses history filtering after switching providers. Preserves readable content and tool-call pairing; no model probes.", "zh": "按供应商开关 Codex Responses 历史过滤,缓解切换上游后的密文和历史 ID 兼容问题;保留正文与工具配对,不运行模型检测。" } }, + { "package": "@magpie-community/middleware-history-trim", "name": "History Trim", "kind": "middleware", "community": true, + "summary": { "en": "Send a request's history smaller: fold a run of blank lines, drop a system prompt written twice and a message repeated right after itself, and cut a past tool result longer than maxToolLength at a word boundary. The turn being sent now is left whole.", "zh": "把请求的历史发得更小:折叠连续空行、去掉逐字重复的 system 提示与紧挨着重复的消息、把历史中超过 maxToolLength 的工具结果在词边界处截断;正在发送的这一轮原样不动。" } }, + { "package": "@magpie-community/agent-agy", "name": "Antigravity CLI", "kind": "agent", "community": true, "icon": "antigravity-color", + "summary": { "en": "Antigravity CLI (agy) on magpie's Agents page: pick one of magpie's models, and the agy in your shells starts on magpie's gateway. Takes the place of the built-in, keeping what it set up.", "zh": "把 Antigravity CLI(agy)加到 magpie 的 Agent 页:选一个 magpie 的模型,终端里的 agy 就从 magpie 的网关启动。接替内置的 agy,保留它原有的设置。" } }, + { "package": "@magpie-community/agent-aider", "name": "Aider", "kind": "agent", "community": true, + "summary": { "en": "Aider (aider.chat) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put its config back.", "zh": "把 Aider(aider.chat)加到 magpie 的 Agent 页:给它选 magpie 的模型,关掉即恢复它原来的配置。" } }, + { "package": "@magpie-community/agent-nanobot", "name": "nanobot", "kind": "agent", "community": true, "icon": "https://raw.githubusercontent.com/HKUDS/nanobot/main/webui/public/brand/nanobot_mark.svg", + "summary": { "en": "nanobot (HKUDS/nanobot) on magpie's Agents page: pick one of magpie's models for it, give it Library MCP servers and skills, and switch it off to put its config back.", "zh": "把 nanobot(HKUDS/nanobot)加到 magpie 的 Agent 页:给它选 magpie 的模型、勾选资源库的 MCP 服务与技能,关掉即恢复它原来的配置。" } }, + { "package": "@magpie-community/agent-reasonix-desktop", "name": "Reasonix 1.x", "kind": "agent", "community": true, "icon": "reasonix-color", + "summary": { "en": "Reasonix 1.x Desktop and CLI on Magpie's models, preserving existing providers. Restart Reasonix after choosing a model; disconnect before uninstalling. Separate from the built-in Reasonix Studio 2.x adapter.", "zh": "为 Reasonix 1.x 桌面版和 CLI 选择 Magpie 模型,保留原有供应商;选模型后重启 Reasonix,卸载前先断开。与内置的 Reasonix Studio 2.x 适配器分开。" } }, + { "package": "@magpie-community/agent-deepzora", "name": "Deepzora", "kind": "agent", "community": true, "icon": "https://raw.githubusercontent.com/SivanCola/deepzora/ad855397cdfdaf2ba7e3e3414583ba777412cbe4/desktop/build/linux/icons/hicolor/128x128/apps/deepzora-desktop.png", + "summary": { "en": "Deepzora Desktop, CLI and ACP on Magpie's models, with a reversible TOML/dotenv bridge that preserves existing providers. Restart Deepzora after choosing a model; disconnect before uninstalling. Disable the standalone Deepzora plugin before switching to this package.", "zh": "为 Deepzora 桌面版、CLI 和 ACP 选择 Magpie 模型,通过可恢复的 TOML/dotenv 桥接保留原有供应商。选模型后重启 Deepzora,卸载前先断开;从独立 Deepzora 插件迁移时先停用旧包。" } }, + { "package": "@magpie-community/agent-pig", "name": "PiG", "kind": "agent", "community": true, + "summary": { "en": "PiG (Pi in Go): add Magpie's endpoint and model catalog, choose models in PiG, and keep startup settings unchanged.", "zh": "为 PiG(Pi in Go)注入 Magpie 端点与模型列表,在 PiG 内选择模型,保持启动设置不变。" } }, + { "package": "@magpie-community/agent-bub", "name": "Bub", "kind": "agent", "community": true, + "summary": { "en": "Bub (bub.build) on magpie's Agents page: pick one of magpie's models and Bub uses it, through ~/.bub/config.yml. Needs Bub 0.5.1 or later.", "zh": "把 Bub(bub.build)加到 magpie 的 Agent 页:给它选一个 magpie 的模型,通过 ~/.bub/config.yml 生效。需要 Bub 0.5.1 或更新版本。" } }, + { "package": "@magpie-community/agent-type4me", "name": "Type4Me", "kind": "agent", "community": true, + "summary": { "en": "Type4Me (macOS voice input) on magpie's Agents page: pick one of magpie's models for its LLM post-processing, used from the next dictation.", "zh": "把 Type4Me(macOS 语音输入)加到 magpie 的 Agent 页:给它的 LLM 后处理选 magpie 的模型,下一次语音输入即生效。" } }, + { "package": "@magpie-community/agent-cowork", "name": "Tuanjie Cowork", "kind": "agent", "community": true, + "summary": { "en": "Tuanjie Cowork (codely) on magpie's Agents page: pick one of magpie's models and Cowork chats with it, on the subscriptions magpie already holds.", "zh": "把团结 Cowork(codely)加到 magpie 的 Agent 页:给它选一个 magpie 的模型,它聊天就走那个模型,用的是 magpie 里已登录的那些订阅的额度。" } }, + { "package": "@magpie-community/agent-opencodereview", "name": "Open Code Review", "kind": "agent", "community": true, + "summary": { "en": "Open Code Review (Alibaba's ocr CLI) on magpie's Agents page: pick one of magpie's models and ocr review / ocr scan use it, read from the config on the next run.", "zh": "把 Open Code Review(阿里巴巴的 ocr CLI)加到 magpie 的 Agent 页:给它选 magpie 的模型,下一次 ocr review / ocr scan 即从配置中读到并生效。" } }, + { "package": "@magpie-community/agent-pi-desktop", "name": "Pi-Desktop", "kind": "agent", "community": true, + "summary": { "en": "Pi-Desktop on magpie's Agents page: magpie keeps a Pi-Desktop plugin folder that adds it as a model provider, with its current models and their thinking levels. Pi-Desktop picks up model changes when it starts.", "zh": "把 Pi-Desktop 加到 magpie 的 Agent 页:magpie 维护一个 Pi-Desktop 插件文件夹,把自己加为模型服务,带上当前模型和推理等级;模型有变化时,Pi-Desktop 重启后生效。" } }, + { "package": "@magpie-community/agent-junie", "name": "Junie CLI", "kind": "agent", "community": true, + "icon": "https://junie.jetbrains.com/favicons/favicon.svg", + "summary": { "en": "An explicit seed copies Magpie's full model catalog into Junie CLI and uses Magpie's selected model as the launch default, without LiteLLM. Native ECO is unavailable for custom profiles in Junie 3893.6.", "zh": "手动 seed 向 Junie CLI 同步全部模型,并将 Magpie 所选模型设为启动默认模型,无需 LiteLLM。Junie 3893.6 的自定义 profile 不支持原生 ECO。" } }, + { "package": "@magpie-community/middleware-reserve", "name": "Reserve", "kind": "middleware", "community": true, + "summary": { "en": "Two routings that keep something in reserve: fill accounts up to a ceiling of their allowance, or use subscriptions before API keys.", "zh": "两种留余量的路由:账号用到额度上限就换下一个,或订阅账号优先于 API key。" } }, + { "package": "@magpie-community/tools-search1api", "name": "Search1API", "icon": "https://s1.dev/icon.png", "kind": "tools", "community": true, + "summary": { "en": "Search1API web search, news, agentic search, page retrieval, sitemaps and trends through Magpie's tools MCP server. Sign in with OAuth or an API key.", "zh": "通过 Magpie 的工具 MCP 服务使用 Search1API 网页搜索、新闻、智能搜索、页面读取、站点链接和趋势;支持 OAuth 或 API key 登录。" } }, + { "package": "@guotree/magpie-websearch", "name": "magpie-websearch", "icon": "https://raw.githubusercontent.com/guotree/magpie-websearch/main/assets/icon-128.png", "kind": "tools", "community": true, + "summary": { "en": "Adds 19 search APIs to Magpie: AnySearch, Bocha, Jina, Kagi, Keenable, Linkup, Metaso, Moonshot, Moonshot CN, Ollama, Parallel, Perplexity, Querit, SerpApi, Serper, StepFun, TinyFish, Z.ai, Zhipu. 4 support keyless access; 15 require keys.", "zh": "为 Magpie 网页搜索增加 19 家搜索源:AnySearch、Bocha、Jina、Kagi、Keenable、Linkup、Metaso、Moonshot、Moonshot CN、Ollama、Parallel、Perplexity、Querit、SerpApi、Serper、StepFun、TinyFish、Z.ai 和智谱。4 家支持免 Key,其余 15 家需要 API Key。" } }, + { "package": "magpie-search-mix", "name": "magpie-search-mix", "kind": "tools", "community": true, + "summary": { "en": "Mixes two search APIs for the AI: more accurate than one. Try DeepSeek + Tavily or You.com.", "zh": "混合搜索:同时问两家搜索 API,比只用一家更准。如 DeepSeek 搭配 Tavily 或 You.com。" } }, + { "package": "magpie-comfyui-link", "name": "ComfyUI", "kind": "tools", "community": true, + "summary": { "en": "A bridge between Magpie agents and locally hosted ComfyUI models, exposing native workflows, model resources and generated images through Magpie's tools MCP endpoint.", "zh": "连接 Magpie Agent 与本地 ComfyUI 模型的桥接层,通过 Magpie 工具 MCP 入口提供原生工作流、模型资源和生成图片。" } }, + { "package": "@magpie-community/tools-specstream", "name": "SpecStream", "kind": "tools", "community": true, + "summary": { "en": "SpecStream web search and page extraction for every agent: answers with the source page of each value, and the text of up to 20 URLs a call. Needs a SpecStream API key.", "zh": "给每个 agent 的 SpecStream 网页搜索和网页提取:每个值都带来源页面,一次最多读取 20 个网址的正文。需要 SpecStream API key。" } }, + { "package": "@magpie-community/tools-calc", "name": "Calc", "kind": "tools", "community": true, + "summary": { "en": "Exact arithmetic, unit conversion and calendar date math as tools for every agent: no network, no key.", "zh": "给每个 agent 的精确计算、单位换算和日历日期计算工具:不联网,不需要 key。" } }, + {"package": "@magpie-community/tools-layerfs", "name": "LayerFS", "kind": "tools", "community": true, "summary": {"en": "LayerFS workspaces for every agent: deduplicated checkpoints and branches behind stable filesystem paths, instead of a fresh copy of the tree for every agent.", "zh": "给每个 agent 的 LayerFS 工作区:稳定的文件系统路径背后是可去重的检查点与分支,而不是每个 agent 各复制一棵项目树。"}}, + { "package": "@magpie-community/page-usage", "name": "Usage Glance", "kind": "page", "community": true, + "summary": { "en": "A page in magpie's window with calls, errors, tokens and cost for today, 7 or 30 days, and the models behind them. Reads only your usage summary.", "zh": "在 magpie 窗口里一眼看到今天、7 天或 30 天的调用、错误、token、费用和对应模型。只读取用量汇总。" } }, + {"package": "@magpie-community/page-amp-usage", "name": "Amp Usage", "kind": "page", "community": true, "icon": "https://raw.githubusercontent.com/magpie-community/plugins/main/packages/amp-usage/page/icon.svg", "summary": {"en": "An Amp usage board: requests, tokens and cache hit rate by hour, day and model, read from your Amp account and legacy local threads. Its background service syncs that usage into Magpie requests and Day card, with duplicate checks; set sync to false to only view it.", "zh": "Amp \u7528\u91cf\u770b\u677f\uff1a\u8bf7\u6c42\u6570\u3001Token\u3001\u7f13\u5b58\u547d\u4e2d\u7387\uff0c\u6309\u5c0f\u65f6\u3001\u6309\u65e5\u548c\u6309\u6a21\u578b\u67e5\u770b\uff0c\u6570\u636e\u6765\u81ea Amp \u8d26\u53f7\u548c\u65e7\u7248\u672c\u5730\u7ebf\u7a0b\u3002\u540e\u53f0\u670d\u52a1\u628a\u8fd9\u4e9b\u7528\u91cf\u540c\u6b65\u8fdb Magpie \u8bf7\u6c42\u7edf\u8ba1\u548c\u6bcf\u65e5\u5361\u7247\uff0c\u5e26\u53bb\u91cd\u68c0\u67e5\uff1b\u628a sync \u8bbe\u4e3a false \u5219\u53ea\u67e5\u770b\u3002"}}, + { "package": "@magpie-community/page-day-card", "name": "Day Card", "kind": "page", "community": true, + "summary": { "en": "A day of your AI usage printed as a receipt: its cost, requests and streak, the day against your average day, each agent and its models, and its tokens. Copy it as a picture or save it as a PNG. Reads only your usage by day.", "zh": "把一天的 AI 用量打印成一张小票:费用、请求数、连续使用天数、与日均的倍数、每个 Agent 和它用的模型,以及 token 用量。可复制为图片或保存为 PNG。只读取按天的用量。" } }, + { "package": "@magpie-community/page-fingerprint", "name": "Fingerprint", "kind": "page", "community": true, + "summary": { "en": "Check for suspected model substitution by comparing response fingerprints across Magpie providers. Uses lm-detector with shared challenges, local evidence history and JSON export; sampling uses provider quota. Reference data is downloaded on demand from lm.ikale.io, falling back to cdn.jsdmirror.com.", "zh": "在 Magpie 内检测响应指纹,为疑似模型替换提供核查线索。基于 lm-detector,支持跨渠道共享挑战、保留本地证据与 JSON 导出;采样会消耗模型额度。参考库按需从 lm.ikale.io 下载,备用源为 cdn.jsdmirror.com。" } }, + { "package": "@magpie-community/page-feishu-sidebar", "name": "Feishu sidebar", "kind": "page", "community": true, + "summary": { "en": "Generate an AppLink to open Magpie Web in the Feishu desktop sidebar. The page uses the Web address you enter and reads no Magpie data.", "zh": "生成飞书链接,在飞书桌面端左侧导航打开 Magpie Web。页面使用你填写的 Web 地址,不读取 Magpie 数据。" } }, + { "package": "@magpie-community/page-chat", "name": "Chat", "kind": "page", "community": true, + "summary": { "en": "A chat page in magpie's window by Magic_zero: talk to magpie's own models with your MCP servers, instructions and skills, images and LaTeX. Runs a local service on 127.0.0.1 that only its own page can use.", "zh": "Magic_zero 做的 magpie 窗口内「对话」页面:用 magpie 自己的模型聊天,带上你的 MCP 服务器、指令和技能,支持图片与 LaTeX。本地服务只在 127.0.0.1,只给自己的页面用。" } }, + { "package": "@magpie-community/page-agent-office", "name": "Agent Office", "kind": "page", "community": true, + "summary": { "en": "Reads the session files of Claude Code, Codex, Copilot CLI, Kimi Code, DeepSeek Harness, Gemini CLI, Qwen Code, Pi, omp, Qoder, WorkBuddy, CodeBuddy and Grok Build on this computer (prompts, replies and tool commands; text can be turned off). Shows them in a real WebGL 3D office with eleven scenes, an isometric/2D office, or a grouped workbench with Today usage, recent conversations and approvals.", "zh": "读取本机 Claude Code、Codex、Copilot CLI、Kimi Code、DeepSeek Harness、Gemini CLI、Qwen Code、Pi、omp、Qoder、WorkBuddy、CodeBuddy、Grok Build 的会话文件(含提问、回复与工具命令,可关闭文本),用真 3D 办公室(十一种场景)、等距/2D 办公室或分组工作台展示,带今日用量、最近对话与批准状态。" } }, + { "package": "@magpie-community/page-usage-tree", "name": "Account usage", "kind": "page", "community": true, + "summary": { "en": "A provider -> account -> model usage tree with per-account calls, errors and cost. Adds the account layer that Usage Glance's usage:summary does not contain, read from Magpie's local usage API.", "zh": "按供应商 -> 账号 -> 模型展示用量,包含每个账号的调用、错误和费用;补充 Usage Glance 的 usage:summary 所没有的账号层,数据来自 Magpie 本地用量 API。" } }, + {"package": "@magpie-community/agent-kimi-desktop", "name": "Kimi", "kind": "agent", "community": true, "summary": {"en": "Kimi desktop Work with Magpie models: configure primary and secondary models on its embedded page, select native subagent roles, and disconnect to restore the original service. Not Kimi Code CLI.", "zh": "普通 Kimi 桌面版 Work 使用 Magpie 模型:在内嵌页面配置主、副模型,为原生子代理选择角色,断开时恢复原服务。不是 Kimi Code CLI。"}}, + {"package": "@magpie-community/agent-minimax-design", "name": "MiniMax Design", "kind": "agent", "community": true, "icon": "https://design.minimax.io/favicon.ico", "summary": {"en": "MiniMax Design (the local multimodal creation studio) on magpie's Agents page: one switch points Design's custom text model at magpie's gateway, and Design's model menu follows magpie's model list automatically. Quit Design before switching; Windows only (see COMPATIBILITY.md).", "zh": "把 MiniMax Design(本地多模态创作工作室)加到 magpie 的 Agent 页:一个开关把 Design 的自定义文本模型指到 magpie 网关,Design 的模型菜单随 magpie 的模型列表自动同步。切换前先退出 Design;仅支持 Windows,详见 COMPATIBILITY.md。"}}, + { "package": "@magpie-community/gateway-notify", "name": "Notify", "kind": "middleware", "community": true, + "summary": { "en": "A push when an agent's turn is done or fails, after it ran long enough to be worth one: ntfy, Bark, Gotify, Server酱, Feishu, DingTalk, WeCom, Telegram or your own webhook. Also pushes when a subscription account is signed out and needs signing in again. Agent, project, model and time only by default; the reply's text is opt-in.", "zh": "agent 一轮任务跑完或失败时推送提醒(只提醒跑够设定时长的任务):支持 ntfy、Bark、Gotify、Server酱、飞书、钉钉、企业微信、Telegram 和自定义 webhook。订阅账号掉线需要重新登录时也会推送。默认只发 agent、项目、模型和用时,回复正文需手动开启。" } }, + { "package": "@magpie-community/hooks-mail-watch", "name": "Mail Watch", "kind": "middleware", "community": true, + "summary": { "en": "Email alerts when a provider call fails or a quota window runs out: your own SMTP account, your own words in the mail, and your own list of what to watch, with cooldowns and an hourly cap so a bad hour stays bounded — plus an optional periodic usage digest.", "zh": "供应商调用失败或额度窗口用尽时发邮件提醒:用你自己的 SMTP 账号、自己的邮件文案、自己勾选的监控事项;同类事件冷却合并,另有每小时上限,糟糕的一小时不会失控;另可开启定期用量摘要。" } }, + { "package": "@magpie-community/opencode-cloudflare-workers-ai-auth", "name": "Cloudflare Workers AI", "providers": ["cloudflare-workers-ai"], "community": true, + "summary": { "en": "Cloudflare Workers AI with your API token (Workers AI:Edit): 28+ models discovered live from models.dev, the 8 paid-only models flagged, and a Neurons/day allowance row in magpie quota.", "zh": "用你的 Cloudflare API Token(Workers AI:Edit 权限)登录 Workers AI:28+ 模型经 models.dev 实时发现,付费专属模型自动标注,magpie quota 显示每日 Neurons 免费额度。" } }, + { "package": "@magpie-community/opencode-openai-decisions-auth", "name": "OpenAI Decisions", "providers": ["openai-decisions"], "community": true, "icon": "data:image/svg+xml,%3Csvg%20xmlns%3D%22http%3A%2F%2Fwww.w3.org%2F2000%2Fsvg%22%20viewBox%3D%220%200%2064%2064%22%3E%3Crect%20width%3D%2264%22%20height%3D%2264%22%20rx%3D%2214%22%20fill%3D%22%23111827%22%2F%3E%3Cpath%20d%3D%22M20%2032h12V20h12M32%2032v12h12%22%20fill%3D%22none%22%20stroke%3D%22white%22%20stroke-width%3D%225%22%20stroke-linecap%3D%22round%22%20stroke-linejoin%3D%22round%22%2F%3E%3Ccircle%20cx%3D%2220%22%20cy%3D%2232%22%20r%3D%225%22%20fill%3D%22white%22%2F%3E%3Ccircle%20cx%3D%2246%22%20cy%3D%2220%22%20r%3D%225%22%20fill%3D%22white%22%2F%3E%3Ccircle%20cx%3D%2246%22%20cy%3D%2244%22%20r%3D%225%22%20fill%3D%22white%22%2F%3E%3C%2Fsvg%3E", "summary": { "en": "OpenAI's native Decisions API as System One for Magpie classifiers and decision groups: choice, score and boolean probabilities, with standard API-key sign-in and per-question refusal handling.", "zh": "将 OpenAI 原生 Decisions API 转为 System One,供 Magpie 分类器与决策模型组使用,保留选择、评分和布尔概率;使用标准 API key 登录,单题拒绝时保留其余答案。" } }, + { "package": "@magpie-community/opencode-openrouter-decisions-auth", "name": "OpenRouter Decisions", "providers": ["openrouter-decisions"], "community": true, "icon": "data:image/svg+xml,%3Csvg%20xmlns%3D%22http%3A%2F%2Fwww.w3.org%2F2000%2Fsvg%22%20viewBox%3D%220%200%2064%2064%22%3E%3Ccircle%20cx%3D%2232%22%20cy%3D%2232%22%20r%3D%2229%22%20fill%3D%22%230f172a%22%2F%3E%3Cpath%20d%3D%22M20%2040a12%2012%200%201%201%2012%2012%22%20fill%3D%22none%22%20stroke%3D%22%2365a3ff%22%20stroke-width%3D%225%22%20stroke-linecap%3D%22round%22%2F%3E%3Ccircle%20cx%3D%2244%22%20cy%3D%2228%22%20r%3D%226%22%20fill%3D%22%2365a3ff%22%2F%3E%3C%2Fsvg%3E", "summary": { "en": "OpenRouter's Decisions API as System One for Magpie classifiers and decision groups: Microsoft Decision 1.1 and 1, with choice, score and boolean probabilities, credit usage and per-question refusal handling.", "zh": "将 OpenRouter Decisions API 转为 System One,供 Magpie 分类器与决策模型组使用:Microsoft Decision 1.1 与 1,保留选择、评分和布尔概率、额度用量,单题拒绝时保留其余答案。" } }, + { "package": "@magpie-community/opencode-river-auth", "name": "River", "providers": ["river"], "community": true, + "summary": { "en": "River (river.ai) pay-as-you-go inference with your API key: per-account models from /v1/models, documented reasoning efforts and live prices from River's public pricing endpoint.", "zh": "用你的 API key 接入 River (river.ai) 按量计费推理:按账号从 /v1/models 列出模型,带官方推理档位与实时价格。" } }, + { "package": "@magpie-community/opencode-sap-aicore-auth", "name": "SAP AI Core", "providers": ["sap-aicore"], "community": true, + "summary": { "en": "Existing SAP AI Core deployments with client-secret credential-file sign-in, per-resource-group GPT Responses and Claude Messages discovery, streaming and agent tools. No deployment provisioning or quota reporting; Magpie only.", "zh": "用本机 client-secret 凭据文件连接已有 SAP AI Core 部署:按资源组发现 GPT Responses 与 Claude Messages 模型,支持流式回复和 agent 工具。不创建部署,不报告额度;仅验证 Magpie。" } }, + { "package": "magpie-plugin-pink", "name": "PinK", "kind": "agent", "icon": "https://raw.githubusercontent.com/YuchenWell/magpie-plugin-pink/main/assets/pink.png", - "summary": { - "en": "Connect PinK to Magpie's models from the Agents page and track gateway usage and cost. Disconnecting restores the managed configuration.", - "zh": "在 Agent 页面连接 PinK,使用 Magpie 中配置的模型并查看调用用量与费用;断开时恢复插件管理的配置。" - } - }, - { - "package": "@hoptoper/magpie-agent-wispterm", - "name": "WispTerm", - "kind": "agent", - "summary": { - "en": "WispTerm on magpie's Agents page: pick one of magpie's models for its Agent, switch reasoning effort with /effort, and switch it off to put its config back. Needs WispTerm 1.46.1 or later.", - "zh": "把 WispTerm 加到 magpie 的 Agent 页:给它的 Agent 选 magpie 的模型,用 /effort 切换思考强度,关掉即恢复原来的配置。需要 WispTerm 1.46.1 或更新版本。" - } - }, - { - "package": "@hoptoper/magpie-agent-citebox", - "name": "CiteBox", - "kind": "agent", - "summary": { - "en": "CiteBox on magpie's Agents page: pick one of magpie's models for its text AI, set reasoning effort in Settings, and switch it off to put its config back. Needs CiteBox 0.38.0 or later.", - "zh": "把 CiteBox 加到 magpie 的 Agent 页:给它的文本 AI 选 magpie 的模型,在设置里调整思考强度,关掉即恢复原来的配置。需要 CiteBox 0.38.0 或更新版本。" - } - }, - { - "package": "@hoptoper/magpie-agent-wisp-science", - "name": "Wisp Science", - "kind": "agent", - "summary": { - "en": "Wisp Science on magpie's Agents page: pick one of magpie's models, set reasoning effort in the composer, and switch it off to put its config back. Needs Wisp Science 1.23.0 or later.", - "zh": "把 Wisp Science 加到 magpie 的 Agent 页:选择 magpie 的模型,在输入框里设置思考强度,关掉即恢复原来的配置。需要 Wisp Science 1.23.0 或更新版本。" - } - }, - { - "package": "magpie-agent-kaku", - "name": "Kaku", - "kind": "agent", + "summary": { "en": "Connect PinK to Magpie's models from the Agents page and track gateway usage and cost. Disconnecting restores the managed configuration.", "zh": "在 Agent 页面连接 PinK,使用 Magpie 中配置的模型并查看调用用量与费用;断开时恢复插件管理的配置。" } }, + { "package": "@hoptoper/magpie-agent-wispterm", "name": "WispTerm", "kind": "agent", + "summary": { "en": "WispTerm on magpie's Agents page: pick one of magpie's models for its Agent, switch reasoning effort with /effort, and switch it off to put its config back. Needs WispTerm 1.46.1 or later.", "zh": "把 WispTerm 加到 magpie 的 Agent 页:给它的 Agent 选 magpie 的模型,用 /effort 切换思考强度,关掉即恢复原来的配置。需要 WispTerm 1.46.1 或更新版本。" } }, + { "package": "@hoptoper/magpie-agent-citebox", "name": "CiteBox", "kind": "agent", + "summary": { "en": "CiteBox on magpie's Agents page: pick one of magpie's models for its text AI, set reasoning effort in Settings, and switch it off to put its config back. Needs CiteBox 0.38.0 or later.", "zh": "把 CiteBox 加到 magpie 的 Agent 页:给它的文本 AI 选 magpie 的模型,在设置里调整思考强度,关掉即恢复原来的配置。需要 CiteBox 0.38.0 或更新版本。" } }, + { "package": "@hoptoper/magpie-agent-wisp-science", "name": "Wisp Science", "kind": "agent", + "summary": { "en": "Wisp Science on magpie's Agents page: pick one of magpie's models, set reasoning effort in the composer, and switch it off to put its config back. Needs Wisp Science 1.23.0 or later.", "zh": "把 Wisp Science 加到 magpie 的 Agent 页:选择 magpie 的模型,在输入框里设置思考强度,关掉即恢复原来的配置。需要 Wisp Science 1.23.0 或更新版本。" } }, + { "package": "magpie-agent-kaku", "name": "Kaku", "kind": "agent", "icon": "https://raw.githubusercontent.com/shlroland/magpie-agent-kaku/main/assets/icon.png", - "summary": { - "en": "Kaku (the terminal) Assistant on magpie's Agents page: one switch points it at magpie's gateway, with a routing group for each of its roles: Simple for # commands and command fixes, Deep for the Cmd+L chat and tools, and optional Memory. Switching it off restores assistant.toml byte for byte. macOS.", - "zh": "把 Kaku 终端的 AI 助手加到 magpie 的 Agent 页:一个开关把它接到 magpie 网关,并给每个角色绑定一个路由分组:Simple 负责 # 命令和报错修复,Deep 负责 Cmd+L 对话和工具调用,Memory 可选。关掉即逐字节恢复 assistant.toml。仅 macOS。" - } - }, - { - "package": "@magpie-community/agent-eta", - "name": "Eta", - "kind": "agent", - "community": true, - "icon": "data:image/svg+xml;base64,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", - "summary": { - "en": "Use Magpie models and routing groups in Eta’s own model picker. Requires Eta 0.1.164 or later; restart Eta after connecting.", - "zh": "在 Eta 自带的模型选择器中使用 Magpie 模型与路由组。需要 Eta 0.1.164 或更新版本;连接后重启 Eta。" - } - }, - { - "package": "@magpie-community/agent-rhino9-ai", - "name": "Magpie-Rhino", - "kind": "agent", - "community": true, - "summary": { - "en": "Magpie-Rhino brings Magpie models, custom groups and strategies into Rhino 9's AI panel through Rhino MCP. Install the companion .yak, restart Rhino, then open its AI panel and choose Magpie; no command required. Local gateway only.", - "zh": "Magpie-Rhino 将模型、自定义组和策略接入 Rhino 9 原生 AI 面板,通过 Rhino MCP 建模。安装配套 .yak 并重启后,直接打开 AI 面板选择 Magpie,无需输入命令。仅连接本机网关。" - } - }, - { - "package": "@magpie-community/tools-goanyapi", - "name": "GoAnyAPI", - "kind": "tools", - "community": true, - "summary": { - "en": "Explore Google Trends, website traffic, domain rating (DR), keyword suggestions, SERPs, backlinks, and ads statistics for market, keyword, and competitor research.", - "zh": "查询 Google 趋势、网站流量、域名权重(DR)、关键词建议、SERP、外链和广告统计,辅助市场、关键词与竞品研究。" - } - }, + "summary": { "en": "Kaku (the terminal) Assistant on magpie's Agents page: one switch points it at magpie's gateway, with a routing group for each of its roles: Simple for # commands and command fixes, Deep for the Cmd+L chat and tools, and optional Memory. Switching it off restores assistant.toml byte for byte. macOS.", "zh": "把 Kaku 终端的 AI 助手加到 magpie 的 Agent 页:一个开关把它接到 magpie 网关,并给每个角色绑定一个路由分组:Simple 负责 # 命令和报错修复,Deep 负责 Cmd+L 对话和工具调用,Memory 可选。关掉即逐字节恢复 assistant.toml。仅 macOS。" } }, + { "package": "@magpie-community/agent-eta", "name": "Eta", "kind": "agent", "community": true, "icon": "data:image/svg+xml;base64,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", + "summary": { "en": "Use Magpie models and routing groups in Eta’s own model picker. Requires Eta 0.1.164 or later; restart Eta after connecting.", "zh": "在 Eta 自带的模型选择器中使用 Magpie 模型与路由组。需要 Eta 0.1.164 或更新版本;连接后重启 Eta。" } }, + { "package": "@magpie-community/agent-rhino9-ai", "name": "Magpie-Rhino", "kind": "agent", "community": true, + "summary": { "en": "Magpie-Rhino brings Magpie models, custom groups and strategies into Rhino 9's AI panel through Rhino MCP. Install the companion .yak, restart Rhino, then open its AI panel and choose Magpie; no command required. Local gateway only.", "zh": "Magpie-Rhino 将模型、自定义组和策略接入 Rhino 9 原生 AI 面板,通过 Rhino MCP 建模。安装配套 .yak 并重启后,直接打开 AI 面板选择 Magpie,无需输入命令。仅连接本机网关。" } }, + { "package": "@magpie-community/tools-goanyapi", "name": "GoAnyAPI", "kind": "tools", "community": true, + "summary": { "en": "Explore Google Trends, website traffic, domain rating (DR), keyword suggestions, SERPs, backlinks, and ads statistics for market, keyword, and competitor research.", "zh": "查询 Google 趋势、网站流量、域名权重(DR)、关键词建议、SERP、外链和广告统计,辅助市场、关键词与竞品研究。" } }, { "package": "@magpie-community/opencode-orcaterm-auth", "name": "OrcaTerm", @@ -1138,17 +212,8 @@ "zh": "用腾讯云遨驰终端(OrcaTerm)AI 助手的模型——DeepSeek、混元、GLM、Kimi——走账号每月问答额度。非官方,条款风险见 README。" } }, - { - "package": "magpie-aws-bedrock-sso", - "name": "AWS Bedrock SSO", - "providers": [ - "bedrock-sso" - ], - "summary": { - "en": "Use Amazon Bedrock Claude and GPT models with your AWS CLI profile, including SSO / IAM Identity Center. SigV4 authentication; no Bedrock API key needed.", - "zh": "使用 AWS CLI 配置(支持 SSO / IAM Identity Center)接入 Amazon Bedrock 的 Claude 和 GPT 模型,通过 SigV4 签名,无需 Bedrock API 密钥。" - } - }, + { "package": "magpie-aws-bedrock-sso", "name": "AWS Bedrock SSO", "providers": ["bedrock-sso"], + "summary": { "en": "Use Amazon Bedrock Claude and GPT models with your AWS CLI profile, including SSO / IAM Identity Center. SigV4 authentication; no Bedrock API key needed.", "zh": "使用 AWS CLI 配置(支持 SSO / IAM Identity Center)接入 Amazon Bedrock 的 Claude 和 GPT 模型,通过 SigV4 签名,无需 Bedrock API 密钥。" } }, { "package": "magpie-providers-hub", "name": "Providers Hub", @@ -1192,90 +257,19 @@ "zh": "窗口预热(原「定时消息」):提前开启订阅的 5 小时或每周额度窗口,给额度保活。给指定的供应商、账号和模型发一条固定短消息(\"hi\",max_tokens 16),可在窗口重置后、每天定时或每隔几小时发送。每次发送都会消耗该供应商的额度(配额)。在本机 127.0.0.1 跑一个本地服务,只调用 magpie 自己的网关;令牌只交给自己的页面。" } }, - { - "package": "@magpie-community/opencode-astudio-auth", - "name": "AStudio", - "providers": [ - "astudio" - ], - "community": true, - "summary": { - "en": "Unofficial AStudio (iFLYTEK Astron Studio): sign in through a browser (each account gets its own pool slot), reuse the desktop sign-in, or paste a token, to serve its Responses model gateway with the live model catalog, daily login-points check-in (read from the credit ledger) and points balance with spark-point expiry. Differs from Providers Hub's hub-astudio, which signs in on a local page (SMS, password or WeChat QR) and calls chat/completions; on one account both share the same daily points pool.", - "zh": "非官方 AStudio(科大讯飞 Astron Studio):支持浏览器登录(每个账号占账号池一格)、复用桌面端登录态或粘贴凭据,接入其模型网关,含实时模型目录、每日登录积分签到(按积分流水判定)与积分余额/星火到期时间。与 Providers Hub 的 hub-astudio 不同:那边在本地登录页用短信、密码或微信扫码登录,走 chat/completions;同一账号两边共用一个每日积分池。" - } - }, - { - "package": "@magpie-community/page-free-tokens", - "name": "Token 羊毛", - "kind": "page", - "community": true, - "summary": { - "en": "AIHOT's Token freebies in magpie's window, in three tiers: already configured and usable, needs one more plugin, or no plugin yet.", - "zh": "AIHOT 的 Token 羊毛,在 magpie 窗口里分三档:已配置可用 / 差一个插件 / 暂无插件。" - } - }, - { - "package": "@magpie-community/opencode-agnes-code-auth", - "name": "Agnes Code", - "providers": [ - "agnes-code" - ], - "community": true, - "summary": { - "en": "Unofficial Agnes Code desktop: reuse the desktop sign-in for its OpenAI-compatible gateway, with the live model catalog, daily credit check-in and points balance. Desktop decryption is Windows-only; a paste-token method covers the rest. Differs from opencode-agnes-auth (`agnes`), which signs in to Agnes AI international in Chrome and meters that product's plans and USD balance: this one serves the Agnes Code desktop product and its daily credits.", - "zh": "非官方 Agnes Code 桌面端:复用桌面端登录态接入其 OpenAI 兼容网关,含实时模型目录、每日积分签到与余额显示。桌面端解密仅限 Windows,其余平台用粘贴 accessToken。与 opencode-agnes-auth(`agnes`)不同:那边在 Chrome 里登录 Agnes AI 国际版、显示该产品的套餐与美元余额;本包服务 Agnes Code 桌面端产品与它的每日积分。" - } - }, - { - "package": "@magpie-community/opencode-tabbit-auth", - "name": "Tabbit", - "icon": "https://web.tabbit.com/favicon.ico", - "providers": [ - "tabbit" - ], - "community": true, - "summary": { - "en": "Unofficial Tabbit (web.tabbit.com, and web.tabbit.ai for international accounts) provider: reuse the Tabbit Browser sign-in on this computer to serve its chat models, with the account's live catalog, cycle usage and the daily check-in. No browser automation: a turn is Tabbit's own room → run → event-stream calls. Client tool calls are not available (Tabbit's tools are its own).", - "zh": "非官方 Tabbit(web.tabbit.com,国际版账号走 web.tabbit.ai)插件:复用本机 Tabbit 浏览器的登录态接入它的聊天模型,含账号实时模型目录、周期用量与每日签到。不走浏览器自动化:一轮对话就是 Tabbit 自己的 建房间 → 发起 → 事件流。不支持客户端工具调用(Tabbit 的工具是它自己的)。" - } - }, - { - "package": "magpie-plugin-codar", - "name": "Codar (科搭)", - "icon": "https://codar.speedai.com/assets/codar-icon.png", - "providers": [ - "codar" - ], - "community": true, - "summary": { - "en": "Codar (科搭), SpeedAI's research desktop app: its 17 models (the auto router plus the DeepSeek, GLM, GPT and Claude families) over OpenAI Chat and Anthropic Messages, with the model list refreshed live from the vendor. Signs in by reading the desktop app's token file on this machine (read-only) or by pasting an access token; the daily check-in reports the claim and carries the next claim time; the card shows general and agent points and any running free-model activity.", - "zh": "科搭(Codar,SpeedAI 出品的研究桌面应用):17 个模型(auto 路由 + DeepSeek、GLM、GPT、Claude 各系),走 OpenAI Chat 与 Anthropic Messages,模型列表从厂商实时刷新。登录可读取本机桌面端令牌(只读)或粘贴访问令牌;每日签到报告领取结果并带上下次可领时间;卡片显示通用与 Agent 点数及进行中的免费模型活动。" - } - }, - { - "package": "@magpie-community/gateway-agentcompass", - "name": "AgentCompass", - "kind": "middleware", - "community": true, - "summary": { - "en": "AgentCompass routing for Magpie: picks : from candidateModels by a task profile's four capability categories, per-effort time and price, and session continuity, orders pool seats by subscription quota, and optionally serves /v1/systemone for Routing Groups.", - "zh": "AgentCompass 路由插件:按任务画像的四类能力需求、各档位的耗时与价格和会话连续性从 candidateModels 选出 :,按订阅配额排号池席位,并可选作为路由组 /v1/systemone 分类器。" - }, - "icon": "data:image/svg+xml;base64,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" - }, - { - "package": "@magpie-community/opencode-mclaw-auth", - "name": "MClaw", - "icon": "https://yun.139.com/w/static/img/LOGO.png", - "providers": [ - "mclaw" - ], - "community": true, - "summary": { - "en": "Unofficial MClaw provider: China Mobile Cloud's MClaw agent (ai.yun.139.com) with an OpenAI-compatible endpoint — the browser sign-in harvests the ticket by itself, auto-fetches the llm key and the token multipliers, serves the live model catalogue and the aitokens pool usage, and renews the ticket before it expires.", - "zh": "非官方 MClaw 插件:中国移动云 MClaw 智能体(ai.yun.139.com),OpenAI 兼容接口——浏览器登录一次自动收票、自动取 llm key 与积分倍率,动态模型目录与 aitokens 池用量,凭证到期前自动续期。" - } - }, + { "package": "@magpie-community/opencode-astudio-auth", "name": "AStudio", "providers": ["astudio"], "community": true, + "summary": { "en": "Unofficial AStudio (iFLYTEK Astron Studio): sign in through a browser (each account gets its own pool slot), reuse the desktop sign-in, or paste a token, to serve its Responses model gateway with the live model catalog, daily login-points check-in (read from the credit ledger) and points balance with spark-point expiry. Differs from Providers Hub's hub-astudio, which signs in on a local page (SMS, password or WeChat QR) and calls chat/completions; on one account both share the same daily points pool.", "zh": "非官方 AStudio(科大讯飞 Astron Studio):支持浏览器登录(每个账号占账号池一格)、复用桌面端登录态或粘贴凭据,接入其模型网关,含实时模型目录、每日登录积分签到(按积分流水判定)与积分余额/星火到期时间。与 Providers Hub 的 hub-astudio 不同:那边在本地登录页用短信、密码或微信扫码登录,走 chat/completions;同一账号两边共用一个每日积分池。" } }, + { "package": "@magpie-community/page-free-tokens", "name": "Token 羊毛", "kind": "page", "community": true, + "summary": { "en": "AIHOT's Token freebies in magpie's window, in three tiers: already configured and usable, needs one more plugin, or no plugin yet.", "zh": "AIHOT 的 Token 羊毛,在 magpie 窗口里分三档:已配置可用 / 差一个插件 / 暂无插件。" } }, + { "package": "@magpie-community/opencode-agnes-code-auth", "name": "Agnes Code", "providers": ["agnes-code"], "community": true, + "summary": { "en": "Unofficial Agnes Code desktop: reuse the desktop sign-in for its OpenAI-compatible gateway, with the live model catalog, daily credit check-in and points balance. Desktop decryption is Windows-only; a paste-token method covers the rest. Differs from opencode-agnes-auth (`agnes`), which signs in to Agnes AI international in Chrome and meters that product's plans and USD balance: this one serves the Agnes Code desktop product and its daily credits.", "zh": "非官方 Agnes Code 桌面端:复用桌面端登录态接入其 OpenAI 兼容网关,含实时模型目录、每日积分签到与余额显示。桌面端解密仅限 Windows,其余平台用粘贴 accessToken。与 opencode-agnes-auth(`agnes`)不同:那边在 Chrome 里登录 Agnes AI 国际版、显示该产品的套餐与美元余额;本包服务 Agnes Code 桌面端产品与它的每日积分。" } }, + { "package": "@magpie-community/opencode-tabbit-auth", "name": "Tabbit", "icon": "https://web.tabbit.com/favicon.ico", "providers": ["tabbit"], "community": true, + "summary": { "en": "Unofficial Tabbit (web.tabbit.com, and web.tabbit.ai for international accounts) provider: reuse the Tabbit Browser sign-in on this computer to serve its chat models, with the account's live catalog, cycle usage and the daily check-in. No browser automation: a turn is Tabbit's own room → run → event-stream calls. Client tool calls are not available (Tabbit's tools are its own).", "zh": "非官方 Tabbit(web.tabbit.com,国际版账号走 web.tabbit.ai)插件:复用本机 Tabbit 浏览器的登录态接入它的聊天模型,含账号实时模型目录、周期用量与每日签到。不走浏览器自动化:一轮对话就是 Tabbit 自己的 建房间 → 发起 → 事件流。不支持客户端工具调用(Tabbit 的工具是它自己的)。" } }, + { "package": "magpie-plugin-codar", "name": "Codar (科搭)", "icon": "https://codar.speedai.com/assets/codar-icon.png", "providers": ["codar"], "community": true, + "summary": { "en": "Codar (科搭), SpeedAI's research desktop app: its 17 models (the auto router plus the DeepSeek, GLM, GPT and Claude families) over OpenAI Chat and Anthropic Messages, with the model list refreshed live from the vendor. Signs in by reading the desktop app's token file on this machine (read-only) or by pasting an access token; the daily check-in reports the claim and carries the next claim time; the card shows general and agent points and any running free-model activity.", "zh": "科搭(Codar,SpeedAI 出品的研究桌面应用):17 个模型(auto 路由 + DeepSeek、GLM、GPT、Claude 各系),走 OpenAI Chat 与 Anthropic Messages,模型列表从厂商实时刷新。登录可读取本机桌面端令牌(只读)或粘贴访问令牌;每日签到报告领取结果并带上下次可领时间;卡片显示通用与 Agent 点数及进行中的免费模型活动。" } }, + {"package": "@magpie-community/gateway-agentcompass", "name": "AgentCompass", "kind": "middleware", "community": true, "summary": {"en": "AgentCompass routing for Magpie: picks : from candidateModels by a task profile's four capability categories, per-effort time and price, and session continuity, orders pool seats by subscription quota, and optionally serves /v1/systemone for Routing Groups.", "zh": "AgentCompass 路由插件:按任务画像的四类能力需求、各档位的耗时与价格和会话连续性从 candidateModels 选出 :,按订阅配额排号池席位,并可选作为路由组 /v1/systemone 分类器。"}, "icon": "data:image/svg+xml;base64,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"}, + { "package": "@magpie-community/opencode-mclaw-auth", "name": "MClaw", "icon": "https://yun.139.com/w/static/img/LOGO.png", "providers": ["mclaw"], "community": true, + "summary": { "en": "Unofficial MClaw provider: China Mobile Cloud's MClaw agent (ai.yun.139.com) with an OpenAI-compatible endpoint — the browser sign-in harvests the ticket by itself, auto-fetches the llm key and the token multipliers, serves the live model catalogue and the aitokens pool usage, and renews the ticket before it expires.", "zh": "非官方 MClaw 插件:中国移动云 MClaw 智能体(ai.yun.139.com),OpenAI 兼容接口——浏览器登录一次自动收票、自动取 llm key 与积分倍率,动态模型目录与 aitokens 池用量,凭证到期前自动续期。" } }, { "package": "@magpie-community/middleware-privacy-gate", "name": "ShyPie Privacy Gate", @@ -1287,160 +281,30 @@ "zh": "由 ShyPie 项目维护。ShyPie 面向 DeepSeek Harness 的隐私优先工作流中 Magpie 一侧的出口门:当 DSH 通过 Magpie 网关使用 DeepSeek 模型时,对 Magpie 无法本地掩码的图片/文件/音频、编码数据块和允许列表之外的模型一律拒绝。它不做文本掩码(文本由 Magpie Privacy 负责);图片和文件默认返回 403,除非 localModels 和 localProviders 都允许。设置 binaryPolicy: guide 可暂扣二进制内容并提示模型选择本地提取工具;结构性检查按失败关闭处理;超过 maxNodes 个值的请求会被拒绝,但超过 scanChars 的文本不会被检查(优先扫描最新内容,每个字段有上限),较长的编码串可能被误判。" } }, - { - "package": "magpie-deepaicode", - "name": "DeepAI Code", - "icon": "https://raw.githubusercontent.com/aqiu817/magpie-deepaicode/main/icon.svg", - "providers": [ - "deepaicode" - ], - "summary": { - "en": "Unofficial DeepAI Code access for magpie: browser sign-in, the desktop app's saved credentials or an API key, direct upstream requests with the device's real machine fingerprint, live models, daily Theos quota and check-in.", - "zh": "非官方接入 DeepAI Code:浏览器登录、桌面端凭据或 API Key,直连上游并使用真实设备指纹,动态模型目录、每日 Theos 额度与签到。" - } - }, - { - "package": "@magpie-community/opencode-agy-cli-auth", - "name": "Antigravity (local agy CLI)", - "icon": "antigravity-color", - "providers": [ - "agy-cli" - ], - "community": true, - "summary": { - "en": "Unofficial provider that runs the agy CLI you installed and signed in to through the Agent Client Protocol (agy acp), serving only the models agy offers this client. Unlike Antigravity ACP, which downloads Google's ACP server and rides the Antigravity IDE's sign-in, it uses your own agy and adds a process pool, a history budget, a bridge for the client's tools, session resume and agy's usage windows.", - "zh": "非官方插件:通过 Agent Client Protocol(agy acp)驱动你自己安装并登录的 agy CLI,只提供 agy 向本客户端开放的模型。与自行下载 Google ACP 服务端、依赖 Antigravity IDE 登录的 Antigravity ACP 不同,它用你本机的 agy,并带进程池、历史体积预算、客户端工具桥接、会话续接与 agy 用量窗口。" - } - }, - { - "package": "opencode-teleagent-auth", - "name": "TeleAgent 模型", - "icon": "https://raw.githubusercontent.com/anbulang/opencode-teleagent-auth/main/assets/icon.svg", - "providers": [ - "teleagent" - ], - "summary": { - "en": "Unofficial TeleAgent (天翼 AI 星辰超级智能体) desktop app models in magpie: the app's own sign-in, no API key, macOS and Windows, with points balance and daily check-in.", - "zh": "非官方接入 TeleAgent(天翼 AI 星辰超级智能体)桌面端:复用桌面端登录态,不需要 API Key,支持 macOS 和 Windows,带积分余额与每日签到。" - } - }, - { - "package": "opencode-xfyun-maas-auth", - "name": "讯飞星辰 MaaS", - "icon": "https://raw.githubusercontent.com/linjicong/opencode-xfyun-maas-auth/main/icon.svg", - "providers": [ - "xfyun-maas", - "xfyun-coding", - "xfyun-token" - ], - "summary": { - "en": "iFlytek Xfyun MaaS: Spark, GLM, DeepSeek, Kimi, Qwen and more — pay-as-you-go, Coding Plan and Token Plan.", - "zh": "讯飞星辰 MaaS:星火、GLM、DeepSeek、Kimi、Qwen 等模型,按量付费与 Coding/Token 套餐。" - } - }, - { - "package": "@magpie-community/page-usage-board", - "name": "Usage Board", - "kind": "page", - "community": true, - "summary": { - "en": "A quota card wall with expiry tiers, daily/per-provider charts, 24h cache-hit and token curves, optional remote attribution and account detail. Runs its own zero-dependency local service reading this machine's Magpie usage data; optional remote attribution queries the configured remote magpie.", - "zh": "额度到期卡片墙 + 每日与 provider 用量图表 + 24 小时缓存命中与 token 曲线,可选远端归因与账号明细。自带零依赖本机服务,读本机 Magpie 用量数据;远端归因会查询所配置的远端 magpie。" - } - }, - { - "package": "magpie-muse-auth", - "name": "Muse (Meta)", - "icon": "https://raw.githubusercontent.com/sonvivu/magpie-muse-auth/main/icon.svg", - "providers": [ - "muse" - ], - "summary": { - "en": "Muse (Meta) subscription for Magpie and OpenCode: device-code sign-in, minted API key, Wonder and Spark models with reasoning efforts. The -contributor models use Meta's Contributor lane (data may be used for training). Unofficial.", - "zh": "使用 Muse (Meta) 订阅:设备码登录,自动换取 API 密钥,提供 Wonder 和 Spark 模型及推理强度设置;以 -contributor 结尾的模型走 Meta 的 Contributor 通道(数据可能用于训练)。(非官方)" - } - }, - { - "package": "@magpie-community/page-model-quota", - "name": "Model Quota", - "kind": "page", - "community": true, - "summary": { - "en": "How many tokens each model has left, merged across every account and provider: each quota window is turned into tokens from your own usage history and split by the model's share in it.", - "zh": "每个模型在所有账号、所有供应商的池子里还剩多少 token:用你自己的用量历史把每个额度窗口换算成 token,再按模型在窗口里的占比摊开。" - } - }, - { - "package": "@magpie-community/page-harnessweight", - "name": "HarnessWeight", - "kind": "page", - "community": true, - "summary": { - "en": "Compare agent harnesses behind your gateway by input tokens, cache, failures, latency and estimated tool-definition size. Its local service reads this machine's usage ledger; a read-only hook samples tool sizes and tracks in-flight requests.", - "zh": "按 Agent 对比网关后面的输入 token、缓存、失败率、延迟与工具定义体积估算。本机服务读取用量记录,只读 hook 采样工具大小并追踪进行中的请求。" - } - }, - { - "package": "@magpie-community/agent-wave", - "name": "Wave", - "kind": "agent", - "community": true, - "icon": "https://github.com/wavetermdev.png", - "summary": { - "en": "Wave Terminal (waveterm) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put its config back.", - "zh": "把 Wave Terminal(waveterm)加到 magpie 的 Agent 页:给它选 magpie 的模型,关掉即恢复它原来的配置。" - } - }, - { - "package": "@magpie-community/agent-zycode", - "name": "zycode", - "kind": "agent", - "community": true, - "summary": { - "en": "zycode (zypi, the VS Code extension, the ACP agent) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put agent.toml back. A host part copies the model into the [model] table of agent.toml, because magpie cannot edit TOML. Restart zycode after a change.", - "zh": "把 zycode(zypi、VS Code 扩展、ACP agent)加到 magpie 的 Agent 页:给它选 magpie 的模型,关掉即恢复 agent.toml。magpie 不能编辑 TOML,所以由宿主部分把模型写进 agent.toml 的 [model] 表。切换后需重启 zycode。" - } - }, - { - "package": "@magpie-community/agent-meowcode", - "name": "MeowCode", - "kind": "agent", - "community": true, - "summary": { - "en": "MeowCode on magpie's Agents page: magpie's models appear in MeowCode's own model list, written to an overlay file MeowCode reads at start-up. Restart MeowCode after choosing a model.", - "zh": "把 MeowCode 加到 magpie 的 Agent 页:magpie 的模型出现在 MeowCode 自己的模型列表里(启动时读取的覆盖层文件)。选择模型后需重启 MeowCode。" - } - }, - { - "package": "@magpie-community/agent-lorca", - "name": "Lorca", - "kind": "agent", - "community": true, - "icon": "https://lorca.app/icon.png", - "summary": { - "en": "Lorca on magpie's Agents page: writes Magpie's gateway, models and MCP servers into Lorca's credentials (~/.lorca/credentials.json). Restart Lorca after connecting.", - "zh": "把 Lorca 加到 magpie 的 Agent 页:将 Magpie 的网关、模型目录与 MCP 工具库写入 Lorca 的凭据配置(~/.lorca/credentials.json)。连接后需重启 Lorca。" - } - }, - { - "package": "magpie-dafeiyu", - "name": "DaFeiYu Desktop Pet", - "kind": "page", - "summary": { - "en": "A guide page for DaFeiYu, a desktop pet that lights a bulb per request and shows quota percentages and API balances. The pet is a separate Windows/macOS app you download yourself; this page does not install or start it. Artwork from PC2005-cloud/dsh-pet, non-commercial use only.", - "zh": "蓝色大肥鱼桌宠的使用指南页:几个请求亮几盏灯,订阅显示额度百分比,API 钱包显示余额。桌宠是需要另外下载的 Windows/macOS 独立 App,本页不会安装或启动它。角色素材来自 PC2005-cloud/dsh-pet,仅限非商用。" - } - }, - { - "package": "@magpie-community/agent-pine", - "name": "Pine", - "kind": "agent", - "community": true, + { "package": "magpie-deepaicode", "name": "DeepAI Code", "icon": "https://raw.githubusercontent.com/aqiu817/magpie-deepaicode/main/icon.svg", "providers": ["deepaicode"], + "summary": { "en": "Unofficial DeepAI Code access for magpie: browser sign-in, the desktop app's saved credentials or an API key, direct upstream requests with the device's real machine fingerprint, live models, daily Theos quota and check-in.", "zh": "非官方接入 DeepAI Code:浏览器登录、桌面端凭据或 API Key,直连上游并使用真实设备指纹,动态模型目录、每日 Theos 额度与签到。" } }, + { "package": "@magpie-community/opencode-agy-cli-auth", "name": "Antigravity (local agy CLI)", "icon": "antigravity-color", "providers": ["agy-cli"], "community": true, + "summary": { "en": "Unofficial provider that runs the agy CLI you installed and signed in to through the Agent Client Protocol (agy acp), serving only the models agy offers this client. Unlike Antigravity ACP, which downloads Google's ACP server and rides the Antigravity IDE's sign-in, it uses your own agy and adds a process pool, a history budget, a bridge for the client's tools, session resume and agy's usage windows.", "zh": "非官方插件:通过 Agent Client Protocol(agy acp)驱动你自己安装并登录的 agy CLI,只提供 agy 向本客户端开放的模型。与自行下载 Google ACP 服务端、依赖 Antigravity IDE 登录的 Antigravity ACP 不同,它用你本机的 agy,并带进程池、历史体积预算、客户端工具桥接、会话续接与 agy 用量窗口。" } }, + { "package": "opencode-teleagent-auth", "name": "TeleAgent 模型", "icon": "https://raw.githubusercontent.com/anbulang/opencode-teleagent-auth/main/assets/icon.svg", "providers": ["teleagent"], + "summary": { "en": "Unofficial TeleAgent (天翼 AI 星辰超级智能体) desktop app models in magpie: the app's own sign-in, no API key, macOS and Windows, with points balance and daily check-in.", "zh": "非官方接入 TeleAgent(天翼 AI 星辰超级智能体)桌面端:复用桌面端登录态,不需要 API Key,支持 macOS 和 Windows,带积分余额与每日签到。" } }, + { "package": "opencode-xfyun-maas-auth", "name": "讯飞星辰 MaaS", "icon": "https://raw.githubusercontent.com/linjicong/opencode-xfyun-maas-auth/main/icon.svg", "providers": ["xfyun-maas", "xfyun-coding", "xfyun-token"], + "summary": { "en": "iFlytek Xfyun MaaS: Spark, GLM, DeepSeek, Kimi, Qwen and more — pay-as-you-go, Coding Plan and Token Plan.", "zh": "讯飞星辰 MaaS:星火、GLM、DeepSeek、Kimi、Qwen 等模型,按量付费与 Coding/Token 套餐。" } }, + { "package": "@magpie-community/page-usage-board", "name": "Usage Board", "kind": "page", "community": true, + "summary": { "en": "A quota card wall with expiry tiers, daily/per-provider charts, 24h cache-hit and token curves, optional remote attribution and account detail. Runs its own zero-dependency local service reading this machine's Magpie usage data; optional remote attribution queries the configured remote magpie.", "zh": "额度到期卡片墙 + 每日与 provider 用量图表 + 24 小时缓存命中与 token 曲线,可选远端归因与账号明细。自带零依赖本机服务,读本机 Magpie 用量数据;远端归因会查询所配置的远端 magpie。" } }, + { "package": "magpie-muse-auth", "name": "Muse (Meta)", "icon": "https://raw.githubusercontent.com/sonvivu/magpie-muse-auth/main/icon.svg", "providers": ["muse"], + "summary": { "en": "Muse (Meta) subscription for Magpie and OpenCode: device-code sign-in, minted API key, Wonder and Spark models with reasoning efforts. The -contributor models use Meta's Contributor lane (data may be used for training). Unofficial.", "zh": "使用 Muse (Meta) 订阅:设备码登录,自动换取 API 密钥,提供 Wonder 和 Spark 模型及推理强度设置;以 -contributor 结尾的模型走 Meta 的 Contributor 通道(数据可能用于训练)。(非官方)" } }, + { "package": "@magpie-community/page-model-quota", "name": "Model Quota", "kind": "page", "community": true, + "summary": { "en": "How many tokens each model has left, merged across every account and provider: each quota window is turned into tokens from your own usage history and split by the model's share in it.", "zh": "每个模型在所有账号、所有供应商的池子里还剩多少 token:用你自己的用量历史把每个额度窗口换算成 token,再按模型在窗口里的占比摊开。" } }, + { "package": "@magpie-community/page-harnessweight", "name": "HarnessWeight", "kind": "page", "community": true, + "summary": { "en": "Compare agent harnesses behind your gateway by input tokens, cache, failures, latency and estimated tool-definition size. Its local service reads this machine's usage ledger; a read-only hook samples tool sizes and tracks in-flight requests.", "zh": "按 Agent 对比网关后面的输入 token、缓存、失败率、延迟与工具定义体积估算。本机服务读取用量记录,只读 hook 采样工具大小并追踪进行中的请求。" } }, + { "package": "@magpie-community/agent-wave", "name": "Wave", "kind": "agent", "community": true, "icon": "https://github.com/wavetermdev.png", "summary": {"en": "Wave Terminal (waveterm) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put its config back.", "zh": "把 Wave Terminal(waveterm)加到 magpie 的 Agent 页:给它选 magpie 的模型,关掉即恢复它原来的配置。"}}, + { "package": "@magpie-community/agent-zycode", "name": "zycode", "kind": "agent", "community": true, "summary": {"en": "zycode (zypi, the VS Code extension, the ACP agent) on magpie's Agents page: pick one of magpie's models for it, and switch it off to put agent.toml back. A host part copies the model into the [model] table of agent.toml, because magpie cannot edit TOML. Restart zycode after a change.", "zh": "把 zycode(zypi、VS Code 扩展、ACP agent)加到 magpie 的 Agent 页:给它选 magpie 的模型,关掉即恢复 agent.toml。magpie 不能编辑 TOML,所以由宿主部分把模型写进 agent.toml 的 [model] 表。切换后需重启 zycode。"}}, + { "package": "@magpie-community/agent-meowcode", "name": "MeowCode", "kind": "agent", "community": true, "summary": {"en": "MeowCode on magpie's Agents page: magpie's models appear in MeowCode's own model list, written to an overlay file MeowCode reads at start-up. Restart MeowCode after choosing a model.", "zh": "把 MeowCode 加到 magpie 的 Agent 页:magpie 的模型出现在 MeowCode 自己的模型列表里(启动时读取的覆盖层文件)。选择模型后需重启 MeowCode。"}}, + { "package": "@magpie-community/agent-lorca", "name": "Lorca", "kind": "agent", "community": true, "icon": "https://lorca.app/icon.png", "summary": {"en": "Lorca on magpie's Agents page: writes Magpie's gateway, models and MCP servers into Lorca's credentials (~/.lorca/credentials.json). Restart Lorca after connecting.", "zh": "把 Lorca 加到 magpie 的 Agent 页:将 Magpie 的网关、模型目录与 MCP 工具库写入 Lorca 的凭据配置(~/.lorca/credentials.json)。连接后需重启 Lorca。"}}, + { "package": "magpie-dafeiyu", "name": "DaFeiYu Desktop Pet", "kind": "page", + "summary": { "en": "A guide page for DaFeiYu, a desktop pet that lights a bulb per request and shows quota percentages and API balances. The pet is a separate Windows/macOS app you download yourself; this page does not install or start it. Artwork from PC2005-cloud/dsh-pet, non-commercial use only.", "zh": "蓝色大肥鱼桌宠的使用指南页:几个请求亮几盏灯,订阅显示额度百分比,API 钱包显示余额。桌宠是需要另外下载的 Windows/macOS 独立 App,本页不会安装或启动它。角色素材来自 PC2005-cloud/dsh-pet,仅限非商用。" } }, + { "package": "@magpie-community/agent-pine", "name": "Pine", "kind": "agent", "community": true, "icon": "https://raw.githubusercontent.com/pmndre/pine/next/docs/images/logo-light.svg", - "summary": { - "en": "Pine (pmndre/pine, the event-driven agent harness on pi's SDK) on magpie's Agents page: pick one of magpie's models and magpie writes its gateway and catalog into ~/.pine/models.json and the model into pine.json, so new conversations start on it. Restart Pine after a change; unofficial.", - "zh": "把 Pine(pmndre/pine,跑在 pi SDK 上的事件驱动 Agent 框架)加到 magpie 的 Agent 页:选一个 magpie 的模型,magpie 把网关和模型目录写进 ~/.pine/models.json、把起始模型写进 pine.json,新对话就用它。换完模型要重启 Pine;非官方适配。" - } - } + "summary": { "en": "Pine (pmndre/pine, the event-driven agent harness on pi's SDK) on magpie's Agents page: pick one of magpie's models and magpie writes its gateway and catalog into ~/.pine/models.json and the model into pine.json, so new conversations start on it. Restart Pine after a change; unofficial.", "zh": "把 Pine(pmndre/pine,跑在 pi SDK 上的事件驱动 Agent 框架)加到 magpie 的 Agent 页:选一个 magpie 的模型,magpie 把网关和模型目录写进 ~/.pine/models.json、把起始模型写进 pine.json,新对话就用它。换完模型要重启 Pine;非官方适配。" } } ] }