Prepare for tinycloud 0.3.18: sites embed v8–v12 — the live-site elements (skill floor → 0.3.18)

.claude-plugin/marketplace.json

@@ -7,14 +7,14 @@
   },
   "metadata": {
     "description": "Deep video and image work for coding agents: analyze, describe, search, caption, clip, detect faces, and publish via the tinycloud CLI.",
-    "version": "0.3.17"
+    "version": "0.3.18"
   },
   "plugins": [
     {
       "name": "tinycloud",
       "source": "./",
       "description": "Video and image understanding and editing via the tinycloud CLI (Cloudglue): watch/see/extract/search/ask/face over videos and images, captions, clips, collections, plus one-command workflows for sales coaching, blog posts, ad analysis, meeting breakdowns, and YouTube publishing.",
-      "version": "0.3.17",
+      "version": "0.3.18",
       "author": { "name": "Cloudglue" },
       "homepage": "https://tinycloud.sh",
       "repository": "https://github.com/cloudglue/tinycloud",

.claude-plugin/plugin.json

@@ -2,7 +2,7 @@
   "name": "tinycloud",
   "displayName": "Tinycloud Video",
   "description": "Video and image understanding and editing via the tinycloud CLI (Cloudglue): watch/see/extract/search/ask/face over videos and images, captions, clips, collections, and one-command video workflows.",
-  "version": "0.3.17",
+  "version": "0.3.18",
   "author": { "name": "Cloudglue", "url": "https://cloudglue.dev" },
   "homepage": "https://tinycloud.sh",
   "repository": "https://github.com/cloudglue/tinycloud",

CLAUDE.md

@@ -233,7 +233,35 @@ schema/list/show/cancel are free, and failed runs auto-refund. `query` is
 also a workflow step node and an agent tool (it joins the LLM tool set,
 unlike `login`). Because the skill teaches the verb, the floor was raised
 to 0.3.17 (same merge-after-CDN gate — the dist PR merges only after CDN
-`channels.stable` = 0.3.17). The host-level `profile` verb and the leading global flags `--home`/`--profile`
+`channels.stable` = 0.3.17). 0.3.18 is a **sites-embed v8–v12 parity and
+publish-guard release** (no new verbs, flags, or feature ids — verbs stay 17,
+features stay 42; the embed features ship in the Cloudglue-served script, so
+already-published sites pick them up without republishing): the embed family
+the binary teaches moved from v7 to v12 — v8 splits delivery into a core
+script plus a lazily-loaded live chunk (`/__cg/embed-live.js`; no
+author-facing change, `/__cg/embed.js` stays the only script tag and
+generators must never add the chunk themselves), v9 opens up the live
+elements (`<cg-chat>` gains `instructions` — a per-site system prompt — and a
+Stop button; `<cg-search>` gains `scope`/`modalities`/`label-filters`/
+`threshold`/`limit`/`group-by="file"`/`sort-by`; `<cg-deep-search>` streams
+its synthesis and takes `exclude-weak-results`; `<cg-face-search>` takes
+`threshold`/`limit` and plays hits as short bounded moments; the JS client's
+`search`/`faceSearch` accept the full v1 `filter` object), v10 makes every
+live panel restylable via `::part()` and emits host events (`cg-results` /
+`cg-answer` / `cg-error`, plus cancelable `cg-resultopen`/`cg-citationopen`
+to override how moments open), v11 adds `<cg-query>` +
+`window.cloudglue.query` (natural-language analytical questions → inline
+table + the compiled SQL — the site-side face of the same engine as the
+`query` verb; point it at an entities/metadata collection), and v12 adds
+`<cg-transcript>` + `window.cloudglue.transcript` (click-to-seek speech
+turns that follow a bound player; unbilled) and `<cg-chapters auto>`
+self-population. The binary-side change: the publish public-site guard now
+also rejects `<cg-query>`, `<cg-transcript>`, and `<cg-chapters auto>` (the
+`auto` attribute makes the otherwise public-safe chapters element
+private-only; explicit-children `<cg-chapters>` stays public-safe). Because
+the skill teaches the new elements and relies on that extended guard, the
+floor was raised to 0.3.18 (same merge-after-CDN gate — the dist PR merges
+only after CDN `channels.stable` = 0.3.18). The host-level `profile` verb and the leading global flags `--home`/`--profile`
 (also `$TINYCLOUD_HOME`; 0.3.3+) relocate state and are intentionally absent
 from `commands --json` — like the launcher's install/update, they're CLI/host
 concerns, not video operations.
@@ -285,9 +313,9 @@ of printing JSON. Any script invoking the binary must redirect `</dev/null`
   metadata sync) vs live-CDN jobs (`Install + smoke` matrix, npx-against-CDN)
   which run only on push to main or manual dispatch — never on PRs, because a
   CDN gap would fail every PR.
-- The live CDN serves 0.3.14 (latest aliases + v-prefixed pinned tarballs
-  for 0.3.0 through 0.3.14, with `manifest.json` + `.sha256`
-  sidecars; `channels.stable` = 0.3.14); all smoke legs are required.
+- The live CDN serves 0.3.17 (latest aliases + v-prefixed pinned tarballs
+  for 0.3.0 through 0.3.17, with `manifest.json` + `.sha256`
+  sidecars; `channels.stable` = 0.3.17); all smoke legs are required.
 - `publish-npm.yml` (tag `v*`): asserts tag == package.json version → gates
   on `generate-manifest.mjs --check` against the live CDN → publishes via
   npm trusted publishing (OIDC, `id-token: write`, npm ≥ 11.5.1 — no token

README.md

@@ -17,11 +17,11 @@ The npm package is a small launcher: on first run it downloads the matching
 platform distribution from Cloudglue's CDN (cached under
 `~/.tinycloud/versions/<version>/`), verifies its checksum, and execs the real
 binary. The package version pins the binary version, so
-`npx @cloudglue/tinycloud@0.3.17` always runs tinycloud 0.3.17. It also adds two
+`npx @cloudglue/tinycloud@0.3.18` always runs tinycloud 0.3.18. It also adds two
 wrapper commands:
 
 ```bash
-tinycloud install --version 0.3.17   # pre-download a version
+tinycloud install --version 0.3.18   # pre-download a version
 tinycloud install --latest          # install latest stable and pin to it
 tinycloud update                    # move to latest stable, prune old versions
 ```
@@ -133,7 +133,7 @@ go to stderr) — pass `--json`.
 | `library` | Build & query Cloudglue collections (create/add/remove/delete) and browse connectors |
 | `jobs` | Poll, wait on, or forget async jobs |
 | `workflow` | Run packaged pipeline recipes (see below) |
-| `publish` | Publish HTML artifacts as Cloudglue Sites; share videos |
+| `publish` | Publish HTML artifacts as Cloudglue Sites (playback embeds everywhere; private sites add live chat/search/query/transcript elements); share videos |
 | `setup` | Configure the Cloudglue API key and service connections |
 | `login` | Browser sign-in — provisions & saves a Cloudglue API key (0.3.10+) |
 

package.json

@@ -1,6 +1,6 @@
 {
   "name": "@cloudglue/tinycloud",
-  "version": "0.3.17",
+  "version": "0.3.18",
   "description": "Agent CLI for deep video and image work, by Cloudglue. Downloads the tinycloud binary on first run.",
   "bin": {
     "tinycloud": "bin/tinycloud.js"

skills/tinycloud/SKILL.md

@@ -265,12 +265,21 @@ Authoring your own recipes: [reference/workflow-authoring.md](reference/workflow
   link (content and playback stay sign-in gated), and platforms cache per exact
   URL, so turning it back off does not retract posted cards. Details in
   [reference/verbs.md](reference/verbs.md).
-- Live-API discovery components (0.3.6+): the same embed script also defines
-  collection-scoped, **private-site-only** elements that let viewers search/chat
-  inside a published site — `<cg-chat>`, `<cg-search>`, `<cg-deep-search>`,
-  `<cg-face-search>` (chat/search/deep-search need a media-descriptions or
-  rich-transcripts collection; face-search needs face-analysis). `publish`
-  **hard-rejects** them on a public site → publish with `--visibility private`.
+- Live-API components (0.3.6+; the fuller v8–v12 surface is taught from
+  0.3.18): the same embed script also defines collection-scoped,
+  **private-site-only** elements that let viewers search/chat/query inside a
+  published site — `<cg-chat>` (optional `instructions` system prompt +
+  Stop), `<cg-search>`/`<cg-deep-search>` (tuning attributes; deep search
+  streams its synthesis), `<cg-face-search>`, `<cg-query>` (analytical
+  questions → inline table + compiled SQL; point at an entities/metadata
+  collection), `<cg-transcript>` (click-to-seek speech turns bound to a
+  player), and `<cg-chapters auto>` (self-populating chapters).
+  Chat/search/deep-search/transcript/chapters-auto need a media-descriptions
+  or rich-transcripts collection; face-search needs face-analysis; query runs
+  over any non-face collection. All are restylable via `::part()` and emit
+  `cg-results`/`cg-answer`/`cg-error` plus cancelable
+  `cg-resultopen`/`cg-citationopen` events. `publish` **hard-rejects** all of
+  them on a public site → publish with `--visibility private`.
   See [reference/verbs.md](reference/verbs.md).
 
 ## 5. Reference (load on demand)

skills/tinycloud/reference/glossary.md

@@ -169,14 +169,25 @@ connector?" or an envelope field needs explaining.
   sites also take `--preview-title` / `--preview-image`). Opting in makes those
   card fields publicly readable — content and playback stay sign-in gated — so
   ask the user first. See reference/verbs.md.
-- **Discovery components (live API, 0.3.6+)** — the same embed script also
-  defines four collection-scoped, **private-site-only** elements that let a
-  viewer search or chat inside a published site and play results inline via
-  `<cg-video>`: `<cg-chat>`, `<cg-search>`, `<cg-deep-search>` (over a
-  media-descriptions / rich-transcripts collection) and `<cg-face-search>` (over
-  a face-analysis collection). They carry no share id, but `tinycloud publish`
-  rejects them on a public site — publish `--visibility private`. See
-  reference/verbs.md.
+- **Live-API components (0.3.6+; v8–v12 surface taught from 0.3.18)** — the
+  same embed script also defines collection-scoped, **private-site-only**
+  elements that let a viewer search, chat, query, or read a transcript inside
+  a published site and play results inline via `<cg-video>`: `<cg-chat>`
+  (optional `instructions` system prompt + Stop), `<cg-search>` (tuning:
+  `scope`/`modalities`/`label-filters`/`threshold`/`limit`/`group-by`/
+  `sort-by`), `<cg-deep-search>` (streamed synthesis;
+  `exclude-weak-results`) — over a media-descriptions / rich-transcripts
+  collection — `<cg-face-search>` (over a face-analysis collection),
+  `<cg-query>` (analytical questions → inline table + compiled SQL; any
+  non-face collection, best on entities/metadata), `<cg-transcript>`
+  (click-to-seek speech turns bound to a player; unbilled), and
+  `<cg-chapters auto>` (self-populating chapters). All restylable via
+  `::part()` with host events (`cg-results`/`cg-answer`/`cg-error`,
+  cancelable `cg-resultopen`/`cg-citationopen`); their implementation loads
+  lazily from `/__cg/embed-live.js` (automatic — never a second script tag).
+  They carry no share id, but `tinycloud publish` rejects them on a public
+  site (0.3.18 extends the guard to query/transcript/chapters-auto) —
+  publish `--visibility private`. See reference/verbs.md.
 
 ## State and isolation (0.3.3+)
 

skills/tinycloud/reference/verbs.md

@@ -654,26 +654,66 @@ private share — directly or through a container — on a public site. The full
 reference ships with the binary as `references/cg-video.md` inside the
 bundled media-artifact skill (under the install's `skills/` directory).
 
-Discovery components (live API, 0.3.6+): the same `/__cg/embed.js` script also
-defines four **collection-scoped** components that let a viewer search or chat
-*inside* a published site and play the referenced moment inline via `<cg-video>`
-— `<cg-chat>` (conversational Q&A, streaming answers + inline moment citations;
-optional `placeholder`), `<cg-search>` (keyword/transcript text search),
-`<cg-deep-search>` (agentic semantic search), and `<cg-face-search>` (upload or
-paste a face image → matching moments). Each takes `collection="<id>"` (plus
-optional `accent-color` and `--cg-height` CSS sizing); the collection's `--type`
-must match the element — `<cg-chat>`/`<cg-search>`/`<cg-deep-search>` need a
-`media-descriptions` or `rich-transcripts` collection, `<cg-face-search>` needs a
-`face-analysis` collection. Unlike the playback elements they carry **no share
-id**, but their live calls are refused on a public site, so `tinycloud publish`
-**hard-rejects** a page that embeds any of them on a public site — publish with
+Live-API components (0.3.6+; expressive surface below is sites embed v8–v12,
+taught from 0.3.18): the same `/__cg/embed.js` script also defines
+**collection-scoped** components that let a viewer search, chat, query, or
+read a transcript *inside* a published site and play the referenced moment
+inline via `<cg-video>`. Their implementation loads lazily from a second
+chunk (`/__cg/embed-live.js`) the moment a page uses one — automatic; never
+emit that file as a script tag, `/__cg/embed.js` stays the only script load.
+
+- `<cg-chat>` — conversational Q&A, streaming answers + inline moment
+  citations, built-in Stop button. Optional `instructions` (the per-site
+  system prompt — persona, tone, scope) and `placeholder`.
+- `<cg-search>` — keyword/transcript text search. Optional tuning: `scope`
+  (`segment` | `file`), `modalities`, `label-filters` (space/comma lists),
+  `threshold`, `limit`, `group-by="file"`, `sort-by` (`score` |
+  `item_count`). Malformed values degrade to defaults, never an error.
+- `<cg-deep-search>` — agentic semantic search; the synthesis streams in and
+  result cards land as found. Optional `scope`, `limit`,
+  `exclude-weak-results`.
+- `<cg-face-search>` — upload or paste a face image → matching moments,
+  played as short bounded clips. Optional `threshold`, `limit`.
+- `<cg-query>` (embed v11) — natural-language analytical question → inline
+  table + row count, with the compiled SQL shown under the answer (the
+  site-side face of the `query` verb's engine). Optional `max-rows`. Point
+  it at an entities or metadata collection — often a **different id** than
+  the one chat/search use on the same page.
+- `<cg-transcript share-id collection [for]>` (embed v12) — click-to-seek
+  speech turns with timestamps + speakers; with `for="<player-id>"` the
+  active turn follows the playhead, without it turns open as modal clips.
+  Unbilled. Height via `--cg-height`.
+- `<cg-chapters auto share-id collection>` (embed v12) — chapters
+  self-populate from the live API instead of baked-in `<cg-chapter>`
+  children; auto chapters always seek-and-continue. The `auto` attribute
+  makes the otherwise public-safe chapters element private-site-only.
+
+Every live panel is restylable from page CSS via `::part()` hooks
+(`panel`/`input`/`send`/`status`/`results`, per-card `result*`; chat adds
+`log`/`message`(+role)/`bubble`/`citation`/`stop`) and emits host events:
+`cg-results`, `cg-answer`, `cg-error` (carries the HTTP status), plus the
+**cancelable** `cg-resultopen`/`cg-citationopen` — `preventDefault()` swaps
+the built-in clip mount for your own handling (embed v10). A
+`window.cloudglue` JS client (responses/search/deepSearch/faceSearch/query/
+transcript/sharesForFiles/renderCitations) backs fully custom UIs.
+
+Each collection-scoped element takes `collection="<id>"` (plus optional
+`accent-color` and `--cg-height` CSS sizing); the collection's `--type` must
+match the element — `<cg-chat>`/`<cg-search>`/`<cg-deep-search>`/
+`<cg-transcript>`/`<cg-chapters auto>` need a `media-descriptions` or
+`rich-transcripts` collection, `<cg-face-search>` needs a `face-analysis`
+collection, `<cg-query>` runs over any non-face collection. Their live calls
+are refused on a public site, so `tinycloud publish` **hard-rejects** a page
+that embeds any of them on a public site (0.3.18 extends the guard to
+`<cg-query>`/`<cg-transcript>`/`<cg-chapters auto>`) — publish with
 `--visibility private`. End to end: build a collection of the right type
 (`library collections create … --type …` → `library collections add` → poll
-`library collections show` until each file is `completed`) → author HTML with
-the component → `tinycloud publish <html> --visibility private`. Full reference
-plus a kitchen-sink page wiring every component (props, JS API, events) ships
-with the binary as `references/cg-video.md` and `references/kitchen-sink.html`
-inside the bundled media-artifact skill.
+`library collections show` until each file is `completed`; for `<cg-query>`,
+`query schema --in collection:<id>` shows what's queryable) → author HTML
+with the component → `tinycloud publish <html> --visibility private`. Full
+reference plus a kitchen-sink page wiring every component (props, JS API,
+events) ships with the binary as `references/cg-video.md` and
+`references/kitchen-sink.html` inside the bundled media-artifact skill.
 
 ### setup — credentials
 

skills/tinycloud/scripts/preflight.sh

@@ -9,7 +9,7 @@ set -u
 
 # Mirror tinycloud-skill.json: min_version / supported_range upper bound
 # (CI diffs these against the manifest).
-MIN_VERSION="0.3.17"
+MIN_VERSION="0.3.18"
 MAX_VERSION_EXCLUSIVE="0.4.0"
 INSTALL_CMD='curl -fsSL https://app.cloudglue.dev/tinycloud.sh | bash'
 # Kept in sync with ../tinycloud-skill.json required_features (CI diffs them).

skills/tinycloud/tinycloud-skill.json

@@ -1,8 +1,8 @@
 {
-  "skill_version": "0.3.17",
+  "skill_version": "0.3.18",
   "tinycloud": {
-    "min_version": "0.3.17",
-    "supported_range": ">=0.3.17 <0.4.0",
+    "min_version": "0.3.18",
+    "supported_range": ">=0.3.18 <0.4.0",
     "required_features": [
       "envelope.v1",
       "watch.v1",