@@ -6,14 +6,10 @@
66
77 <p align="center"><strong>Lightning-fast coding agent for your terminal.</strong></p>
88
9-<p align="center">
10- Small, fast, focused, and just opinionated enough to be useful.
11-</p>
12-
139 <p align="center">
1410 <a href="https://www.npmjs.com/package/mini-coder">npm</a>
1511 ·
16- <a href="https://sacenox.github.io/mini-coder/">gh-pages</a>
12+ <a href="https://sacenox.github.io/mini-coder/">docs</a>
1713 ·
1814 <a href="spec.md">spec</a>
1915 </p>
@@ -24,83 +20,73 @@
2420 </picture>
2521 </p>
2622
27-mini-coder (`mc`) is a terminal coding agent that reads your repo, edits files, runs commands, and keeps going until the work is done. The rewrite trims the product down to the essentials: a sharp system prompt, a small tool surface, streamed terminal UX, and strong foundations instead of framework soup.
28-
29-## Why mini-coder?
30-
31-- **Terminal-native** — no browser tab, no web app shell, no waiting around.
32-- **Tiny core tool surface** — `shell` and `edit` do the real work. `readImage` joins when the model supports vision.
33-- **Streams everything** — assistant text, reasoning, tool calls, and tool output show up as they happen.
34-- **Actually remembers things** — sessions are persisted in SQLite with undo, fork, resume, and cumulative usage stats.
35-- **Plays well with the community** — supports [AGENTS.md](https://agents.md), [Agent Skills](https://agentskills.io), and plugins.
36-- **Built on proven parts** — [pi-ai](https://github.com/badlogic/pi-mono/tree/main/packages/ai) for providers and [cel-tui](https://github.com/sacenox/cel-tui) for the TUI.
37-
38-## Install the stable release
23+mini-coder (`mc`) is a terminal coding agent that reads your repo, edits files, runs commands, and keeps going until the work is done. Small core tool surface, flat architecture, fast turns, and streaming everywhere it matters.
3924
40-The published npm release installs a working `mc` binary:
25+## Install
4126
4227 ```bash
43-npm install -g mini-coder
44-mc
28+$ bun add -g mini-coder
29+$ mc
4530 ```
4631
47-If you prefer Bun globally:
48-
49-```bash
50-bun add -g mini-coder
51-mc
52-```
53-
54-## Try the rewrite (`0.5.x`)
55-
56-The rewrite is in progress on this repo right now. If you want the new version before it is published, clone the repo and run it directly with Bun:
57-
58-```bash
59-git clone git@github.com:sacenox/mini-coder.git
60-cd mini-coder
61-bun install
62-bun run src/index.ts
63-```
64-
65-Or give yourself a handy alias:
66-
67-```bash
68-alias mc-dev='bun run ~/src/mini-coder/src/index.ts'
69-mc-dev
70-```
71-
72-## What you get
73-
74-- **Coding-agent workflow** — inspect with shell, mutate with edit, verify with shell.
75-- **Multi-provider model support** — Anthropic, OpenAI, Google, OpenRouter, Ollama, Copilot, and more via pi-ai.
76-- **Streaming TUI** — markdown conversation log, tool blocks, animated divider, a two-line default input, and a one-line pill status bar.
77-- **Session persistence** — resume old sessions, `/fork` them, `/undo` turns, and keep working.
78-- **Reasoning + verbosity controls** — toggle thinking visibility and full tool output on demand.
79-- **Prompt context from your repo** — AGENTS.md discovery, skill catalog support, git state in the prompt footer.
80-- **Plugin path** — optional tools, integrations, theme overrides, and prompt suffixes without bloating the core.
81-
82-## Handy commands
83-
84-- `/model` — switch models
85-- `/session` — resume a session
86-- `/new` — start fresh
87-- `/fork` — branch the conversation
88-- `/undo` — remove the last conversational turn
89-- `/reasoning` — toggle thinking visibility
90-- `/verbose` — toggle full tool output
91-- `/login` / `/logout` — manage OAuth providers
92-- `/effort` — set reasoning effort
93-- `/help` — list the whole menu
94-
95-## Status
96-
97-mini-coder has a stable published release on npm today.
32+## Why mini-coder?
9833
99-The rewrite tracked in this repo is the next generation: smaller, faster, more spec-driven, and more focused on being a great coding agent in the terminal. If you want the stable packaged release, install from npm. If you want the shiny new hotness, run `main` from source.
34+- **Lean on proven dependencies** — [pi-ai](https://github.com/badlogic/pi-mono/tree/main/packages/ai) for providers, streaming, tool calling, usage tracking, and OAuth. [cel-tui](https://github.com/sacenox/cel-tui) for the terminal UI. The core stays focused on agent work.
35+- **Flat, simple codebase** — no workspaces, no internal abstraction layers. Files grouped by concern in a single `src/` directory.
36+- **Agent-first** — every decision serves the goal of reading code, making changes, and verifying them via the shell.
37+- **Performance** — startup and turn latency matter more than features.
38+- **Streaming end-to-end** — assistant text, reasoning, tool calls, and tool output show up as they happen.
39+
40+## Tools
41+
42+Two built-in tools, plus a read-only image tool:
43+
44+- **`shell`** — runs commands in the user's shell. Returns stdout, stderr, and exit code. Large output is truncated to protect model context.
45+- **`edit`** — exact-text replacement in a single file. Fails deterministically if the target is missing or ambiguous. Creates new files when old text is empty.
46+- **`readImage`** — reads PNG, JPEG, GIF, and WebP files as model input. Only registered when the active model supports images.
47+
48+Plugins can add more tools, but the core stays intentionally small.
49+
50+## Features
51+
52+- **Multi-provider model support** — Anthropic, OpenAI, Google, Bedrock, Mistral, Groq, xAI, OpenRouter, Ollama, Copilot, and more via pi-ai.
53+- **Streaming TUI** — markdown conversation log, tool blocks with diffs, animated divider, multi-line input, and a one-line pill status bar.
54+- **Session persistence** — SQLite-backed sessions with undo, fork, resume, and cumulative usage stats. Sessions are scoped to the working directory.
55+- **Reasoning and verbosity controls** — toggle thinking visibility and full tool output on demand. Preferences persist across launches.
56+- **[AGENTS.md](https://agents.md) support** — project-specific instructions discovered root-to-leaf, with `~/.agents/` for global instructions.
57+- **[Agent Skills](https://agentskills.io)** — skill catalogs exposed in the prompt. `/skill:name` injects a skill body into the next user message.
58+- **Plugins** — optional tools, integrations, theme overrides, and prompt suffixes without bloating the core.
59+
60+## Commands
61+
62+| Command | Description |
63+| ------------ | ---------------------------------------------------------------------- |
64+| `/model` | Switch models. Persists as the global default. |
65+| `/session` | List and resume sessions scoped to the current working directory. |
66+| `/new` | Start a fresh session and reset cumulative usage counters. |
67+| `/fork` | Copy the conversation into a new session and continue independently. |
68+| `/undo` | Remove the last conversational turn (does not revert file changes). |
69+| `/reasoning` | Toggle thinking visibility. Persisted and restored on launch. |
70+| `/verbose` | Toggle full shell output and edit diff display. |
71+| `/login` | Interactive OAuth login for supported providers. |
72+| `/logout` | Clear saved OAuth credentials for a logged-in provider. |
73+| `/effort` | Set reasoning effort: low, med, high, or xhigh. |
74+| `/help` | List commands, loaded AGENTS.md files, discovered skills, and plugins. |
75+
76+## Key bindings
77+
78+| Key | Action |
79+| ------------- | ------------------------------------------------- |
80+| `Enter` | Submit message |
81+| `Shift+Enter` | Insert newline |
82+| `Escape` | Interrupt current turn, preserve partial response |
83+| `Tab` | File path autocomplete (or command filter on `/`) |
84+| `Ctrl+R` | Search global raw input history |
85+| `Ctrl+C` | Graceful exit |
10086
10187 ## Docs
10288
103-- **Product site:** https://sacenox.github.io/mini-coder/
89+- **Docs site:** https://sacenox.github.io/mini-coder/
10490 - **Spec:** [`spec.md`](spec.md)
10591 - **Repo instructions:** [`AGENTS.md`](AGENTS.md)