Bring docs and manifests in line with the current code

.gitignore

@@ -1,4 +1,4 @@
 /target
 plans
-sum*
+/sum*
 Cargo.toml.orig

AGENTS.md

@@ -2,13 +2,16 @@
 
 ## Project
 
-Alan is minimal coding-agent written in rust.
+Alan is a minimal coding agent written in Rust.
 
 Workspace crates:
 
 - `crates/llm/` — provider-independent LLM protocol types and API clients.
 - `crates/providers/` — provider/model binding and authentication.
+- `crates/tools/` — built-in file system and shell tools.
 - `crates/agent/` — conversation state, tool calls, system prompt, agent loop.
+- `crates/tui/` — package `alan-tui`, a standalone Ratatui runtime: components,
+  event loop, terminal lifecycle. Published to crates.io.
 - `crates/alan/` — interactive ratatui REPL frontend.
 
 Workspace manifest: `Cargo.toml`.
@@ -36,15 +39,17 @@ cargo check -p alan
 cargo test -p agent
 cargo test -p llm
 cargo test -p providers
+cargo test -p tools
+cargo test -p alan-tui
 ```
 
 Run formatting and checks after Rust code changes. Do not run release or destructive Git commands unless requested.
 
 ### Model selection
 
-`/models` opens a picker backed by the OpenRouter model catalog. The catalog
-is fetched on first use and cached; selecting a model switches the active
-conversation model and updates the session header.
+`/models` opens a picker over the model catalogs of every registered provider.
+Selecting a model switches the active conversation model and updates the
+session header.
 
 Profiles save the current provider/model, reasoning effort, and web-fetch/search
 settings. Use `/profile save <name>` to create one, `/profile` to switch, and
@@ -53,7 +58,9 @@ settings for future launches; manual settings changes clear its active-profile
 marker. Profiles are stored separately in `<data dir>/profiles.json`. See
 `core::paths` for how the data directory is resolved.
 
-Alan currently uses OpenRouter:
+`build_providers` in `crates/alan/src/main.rs` lists the registered providers.
+The `providers` crate also has a Google provider that is not registered. Sign in
+with `/login`, or set the provider's API key variable:
 
 ```bash
 OPENROUTER_API_KEY=... cargo run -p alan
@@ -65,36 +72,49 @@ Optional model override:
 ALAN_MODEL=openai/gpt-4o-mini cargo run -p alan
 ```
 
-Current default model is `openai/gpt-4o-mini`.
+With no saved settings, Alan uses the `openrouter` provider and
+`DEFAULT_MODEL` from `crates/alan/src/core/settings.rs`.
 
 ## Architecture
 
 Keep dependency direction one-way:
 
 ```text
 llm <- providers <- agent <- alan
+llm <- tools     <- agent
+alan-tui               <- alan
 ```
 
 `llm` must not depend on providers, agent, or UI.
 
 `agent` must not depend on ratatui or crossterm.
 
-`alan` owns terminal setup, key mapping, and ratatui rendering.
+`alan-tui` must not depend on other workspace crates.
+
+`alan-tui` owns terminal setup and the event loop. `alan` owns key mapping and
+ratatui rendering.
 
 ### Alan frontend boundary
 
 ```text
 crates/alan/src/main.rs
-  terminal lifecycle and crossterm event polling
+  provider setup, model selection, runtime launch
+
+crates/alan/src/keymap.rs
+  crossterm events to `AlanAction`
+
+crates/alan/src/root.rs
+  root component: routes `AlanAction` to children, lays out the screen
 
 crates/alan/src/core/
-  UI-independent actions, controller, transcript, agent execution
+  UI-independent chat controller, transcript, commands, settings,
+  permissions, skills
 
 crates/alan/src/views/
-  ratatui adapter, UI state, layout, rendering
+  ratatui components and theme
 ```
 
-`core` must stay frontend-independent. Future GPUI frontend should reuse `Controller`, `Entry`, and `Action`, then provide its own event mapping and renderer.
+`core` must stay frontend-independent. A future GPUI frontend should reuse `ChatController` and `Entry`, then provide its own event mapping and renderer.
 
 Do not add ratatui/crossterm types to `crates/alan/src/core/`.
 
@@ -111,7 +131,7 @@ auto-includes it in release archives.
 
 ## Current Alan UI
 
-`crates/alan/src/views/mod.rs` provides minimal borderless ratatui UI:
+`crates/alan/src/views/components/` provides minimal borderless ratatui UI:
 
 - Header.
 - Scrollable chat history.
@@ -121,27 +141,15 @@ auto-includes it in release archives.
 - Status line above editor.
 - Input cursor placement.
 
-`UiState` owns frontend interaction state:
-
-- input text
-- transcript scroll offset
-- auto-follow output state
-
-`Action` is UI-independent input intent:
-
-- `Quit`
-- `Submit`
-- `ClearInput`
-- `Backspace`
-- `Insert(char)`
-- `ScrollUp`
-- `ScrollDown`
+`ChatView` (`views/components/chat_view/`) owns the chat session: the
+`ChatController` and the agent stream. `PromptEditor` owns input. Overlays
+such as login and local models live beside `root.rs`.
 
 Keep UI simple. Avoid borders, unnecessary widgets, and premature abstraction.
 
 ## Current Agent Behavior
 
-`crates/agent/src/agent.rs` currently:
+`crates/agent/src/agent/` currently:
 
 - Stores conversation history in memory.
 - Sends prompts through bound `providers::Model`.
@@ -198,11 +206,9 @@ skip is `tracing::warn!`-logged.
 ## UI Rules
 
 - No core dependency on terminal framework.
-- Frontend converts native events into `Action`.
+- Frontend converts native events into `AlanAction`.
 - Renderer reads state; it should not perform network calls.
-- Controller owns agent interaction and transcript state.
-- Keep visual constants centralized in view module.
+- `ChatController` owns agent interaction and transcript state.
+- Keep visual constants centralized in `views/theme.rs`.
 - Use terminal display width for cursor/layout calculations, not byte count.
 - Keep auto-scroll behavior explicit.
-
-Implement in this order unless user requests different scope. Keep each step usable.

CHANGELOG.md

@@ -26,7 +26,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Popup v2 clears before rendering a message such that background text doesn't overlap with popup's message.
 - Text selection in the chat history is precise now.
 
-## 0.1.4 - 2026-10-4
+## 0.1.4 - 2026-10-04
 
 ### Added
 

README.md

@@ -21,7 +21,7 @@ A minimal coding agent in your terminal.
 
 ## Features
 
-- **First-class OpenRouter support** — full model catalog, per-model custom provider pinning (`/model-provider`), and reasoning effort control (`/effort`).
+- **First-class OpenRouter support** — full model catalog, per-model custom provider pinning (`/model-providers`), and reasoning effort control (`/effort`).
 - **Paste images** — drop a screenshot from your clipboard straight into the conversation and ask about it.
 - **Compact with focus** — `/summarize-new "focus hint"` summarizes the context and restarts with just what matters.
 - **Steering** — queue a message mid-run and the agent picks it up without waiting for the response to finish.
@@ -75,10 +75,10 @@ Alan currently supports:
 
 Use the `/login` command to sign in to any of the available providers.
 
-For OpenRouter, you can pin a custom provider for the selected model with `/model-provider`, followed by the provider name (no quotes):
+For OpenRouter, you can pin a custom provider for the selected model with `/model-providers`, followed by the provider name (no quotes):
 
-- `/model-provider deepseek`
-- `/model-provider xiaomi/fp8`
+- `/model-providers deepseek`
+- `/model-providers xiaomi/fp8`
 
 ## Usage
 
@@ -96,12 +96,14 @@ For OpenRouter, you can pin a custom provider for the selected model with `/mode
 | `/summarize-new [focus]` | Summarize the context and restart; optionally pass a quoted focus hint, e.g. `/summarize-new "just take the XYZ details"` |
 | `/effort` | Set reasoning effort (`none`, `minimal`, `low`, `medium`, `high`, `xhigh`, `max`) |
 | `/fork` | Fork the session from a checkpoint |
+| `/rename <name>` | Rename the current session |
 | `/tool-free` | Allow all tool calls without asking |
 | `/tool-slip` | Approve a command family once, allow its siblings |
 | `/tool-strict` | Allow only exact commands already approved |
 | `/local` | Manage local models (add, remove, edit) |
-| `/model-provider <name>` | Pin a custom provider for the current OpenRouter model |
+| `/model-providers <name>` | Pin a custom provider for the current OpenRouter model |
 | `/help` | List available commands |
+| `/quit` | Exit Alan |
 
 ## Permissions
 
@@ -115,9 +117,10 @@ A single tool call may chain several commands (`cargo test && cargo run`, or wit
 
 Grants persist per-project at `<data dir>/projects/<pwd-hash>/permissions.json`. The current policy is shown as a glyph in the status line.
 
-Alan starts in strict policy. Switch with:
+Alan starts in strict policy and remembers the policy you pick across runs. Switch with:
   - `/tool-slip`
   - `/tool-free`
+  - `/tool-strict`
 
 ## Local models
 
@@ -127,22 +130,46 @@ Alan supports OpenAI-compatible local models.
 /local add
 ```
 
-This opens an overlay with fields for Model ID, URL, API, and API Key. Use `/local remove` and `/local edit` to manage your local models. No API key or `/login` is needed for local models.
+This opens an overlay with fields for URL, API, API Key, and Model ID. Alan queries the server's `/models` endpoint and offers the detected models in a searchable picker. Use `/local remove` and `/local edit` to manage your local models. No API key or `/login` is needed for local models.
 
 ## Configuration and data locations
 
 Alan stores everything under `~/.alan/` by default:
 
 - `settings.json` — your current settings and active-profile marker.
 - `profiles.json` — saved named model/reasoning/web settings profiles.
+- `auth.json` — provider credentials saved by `/login`.
+- `local_models.json` — local models added with `/local`.
+- `projects/` — per-project permission grants.
 - `sessions/` — conversation history (append-only JSONL, one file per session).
 - `skills/` — personal skills, available in every project.
 - `logs/` — daily rotating logs.
 
-All of these live in the data directory (`~/.alan` by default). `ALAN_HOME` selects the *parent* of that directory, so `ALAN_HOME=/tmp/x` puts everything in `/tmp/x/.alan`. Two other environment variables are useful:
+All of these live in the data directory (`~/.alan` by default).
 
-- `ALAN_LOG_DIR` — override the log directory.
-- `ALAN_MODEL` — override the default model on startup, e.g. `ALAN_MODEL=openai/gpt-4o-mini alan`.
+### Environment variables
+
+Settings variables override `settings.json` for that run without changing it. Start Alan with `--save` to write them into `settings.json`.
+
+| Variable | Effect |
+| --- | --- |
+| `ALAN_HOME` | Parent of the data directory: `ALAN_HOME=/tmp/x` puts everything in `/tmp/x/.alan`. |
+| `ALAN_PROVIDER` | Provider id: `openrouter`, `deepseek`, `zai`, or `local`. |
+| `ALAN_MODEL` | Model id for that provider, e.g. `ALAN_MODEL=openai/gpt-4o-mini alan`. |
+| `ALAN_REASONING_EFFORT` | `none`, `minimal`, `low`, `medium`, `high`, `xhigh`, or `max`. |
+| `ALAN_OPENROUTER_WEB_FETCH` | Turn OpenRouter's web fetch tool on or off: `true`/`false`, `1`/`0`, `yes`/`no`, `on`/`off`. |
+| `ALAN_OPENROUTER_WEB_SEARCH` | Turn OpenRouter's web search tool on or off, same values. |
+| `ALAN_OR_MODEL_PROVIDER` | Comma-separated OpenRouter provider order for the selected model, e.g. `deepseek,xiaomi/fp8`. |
+| `ALAN_SESSION` | Resume a saved session by id. Alan prints the exact command when it exits. |
+| `ALAN_LOG` | Log filter in `tracing` `EnvFilter` syntax. Falls back to `RUST_LOG`. |
+| `ALAN_LOG_DIR` | Log directory, default `~/.alan/logs`. |
+| `OPENROUTER_API_KEY`, `DEEPSEEK_API_KEY`, `ZAI_API_KEY` | API key for that provider, used when `/login` has not saved one. |
+
+### Flags
+
+- `--save` — write the settings environment variables above into `settings.json`.
+- `--blank` — start without Alan's default system prompt.
+- `--version` — print the installed version.
 
 ## Skills
 

crates/alan/Cargo.toml

@@ -38,11 +38,9 @@ tracing-subscriber = { workspace = true }
 tracing = { workspace = true }
 tracing-appender = { workspace = true }
 ignore = { workspace = true }
-fs2 = "0.4"
+fs2 = { workspace = true }
 
 [dev-dependencies]
-llm = { workspace = true }
-async-trait = { workspace = true }
 tokio = { workspace = true, features = ["test-util"] }
 
 [[bin]]

crates/llm/Cargo.toml

@@ -12,6 +12,3 @@ serde = { workspace = true }
 serde_json = { workspace = true }
 thiserror = { workspace = true }
 tokio = { workspace = true }
-tracing-subscriber = { workspace = true }
-tracing = { workspace = true }
-tracing-appender = { workspace = true }

crates/providers/Cargo.toml

@@ -13,5 +13,4 @@ reqwest = { workspace = true }
 tokio = { workspace = true, features = ["sync", "rt"] }
 
 [dev-dependencies]
-futures-util = { workspace = true }
 tokio = { workspace = true }

crates/tui/examples/greet.rs

@@ -2,7 +2,7 @@
 //! greets you and echoes what you typed.
 //!
 //! The editor is configured the same way alan configures its prompt
-//! (`UiState::new_editor`): word wrap, hidden in-widget cursor, and the real
+//! (`PromptEditor::new_editor`): word wrap, hidden in-widget cursor, and the real
 //! terminal cursor positioned from `rendered_cursor_position`.
 
 use std::time::Duration;
@@ -24,7 +24,7 @@ struct Greet {
 }
 
 impl Greet {
-    /// Same configuration as alan's `UiState::new_editor`.
+    /// Same configuration as alan's `PromptEditor::new_editor`.
     fn new_editor() -> TextArea<'static> {
         let mut editor = TextArea::default();
         editor.set_style(Style::default());