Monet

Multi-engine mission control for coding agents — Claude Code and Codex already seated

Features • Install • Build • Privacy • FAQ

Don't feel like reading? Hand this line to your AI — it will assess, install, and troubleshoot for you:
Read https://raw.githubusercontent.com/zenolab124/monet/main/llms.txt
_{If you are that AI: read the llms.txt above and follow its guidance to serve your user.}

Monet race workbench — four agents creating in parallel with live result comparison

## What is Monet? Your conversations with coding agents are scattered across terminals, projects, and session histories — reviewing means digging, parallel work means switching windows, and long-running jobs keep pulling you back to check progress. Monet gathers Claude Code, Codex, and future engines onto one wall: every agent session browsable, searchable, and commandable in parallel. The agent does the work; Monet gives you eyes and hands. ## Why Monet? **However many engines, one wall.** Claude Code and Codex share one Archive, one search, and one Workbench, with an engine badge on every session. Race mode pits engines against the same prompt, while scheduled routines can run on whichever engine you choose — selecting an engine feels as natural as selecting a model. **Command your agents like a trader watches the market.** Unlimited parallel session columns spread out horizontally; the monitor rail puts status, output, and token usage at a glance. Approve permissions, answer questions, and retry failures directly from the cards. **A home for multi-channel players.** Official subscriptions, third-party APIs, self-hosted proxies, local models — each session can use its own channel and switch mid-conversation. Put a strong model on the hard problems and a cheaper channel on routine work, with connections and models always visible. **Your data stays yours.** Claude Code JSONL and Codex rollouts are read-only by architecture, while Monet's titles, tags, stars, and other metadata live separately in `~/.monet/`. Local session features work offline, with zero telemetry and no Monet account system; only subscription quota and AI enhancements you explicitly use contact the relevant provider services. ## Features ### Multi-engine system — Claude Code and Codex together - Claude Code and Codex projects, sessions, and timelines enter the same Archive, search, Workbench, and notifications, with filters by engine - You can browse existing history without the corresponding CLI; install and sign in to the CLI to create, resume, and keep running sessions in Monet - Both engines share a familiar session shell and input experience, while the interface only shows actions the current engine actually supports - Engine Center reports installation, authentication, version, capabilities, and diagnostics independently; one broken engine does not block another, and future engines can join through the same contract ### Workbench — parallel agent command - No cap on column count; when the screen runs out, scroll horizontally — wheel input glides like a native trackpad - Monitor rail overviews every session: live status, tail output, token usage; approve/retry/answer right on the card - Permission requests as GUI cards: dangerous commands flagged in red, AI annotates the risk in plain language; `Enter` to allow, `Esc` to deny - **Race mode**: broadcast one question to different engines, models, or channels, then compare answers and cost side by side - **Panorama export**: capture the entire workbench wall as one image — share or archive the battle at a glance

Monet multi-column Workbench — session monitoring, parallel columns, inline HTML rendering, and live background tasks

### Session running — the CLI session, now in a GUI - True character-level streaming; first-token latency down to the API's first token - **Hot-switch channel / model / thinking effort mid-session** — no process restart, no lost context - File ledger: which files this session touched, every diff, one click back to "why the AI changed it" - Runner supervises your dev server: logs right next to the session, errors fed to the AI in one click - Sessions running in your terminal are auto-detected and followed live — you can even terminate them from Monet ### Reading experience — one engine for live and past - Three tool-call display modes: full cards, collapsed items, or grouped consecutive processes; expanding preserves purpose-built features such as side-by-side Edit diffs and separately copyable Bash sections - Inline HTML/SVG rendering in replies: comparison cards, tables, diagrams — no more walls of text; LaTeX math rendered natively - **Artifact previews in place**: local HTML, SVG, and image files linked in a reply become preview cards automatically — click to zoom, zero configuration; HTML loads statically in a script-free, network-free sandbox by default - Anchor navigation + pinned prompts + back-to-bottom float: glide through sessions hundreds of turns long - Per-turn token quadruple and a context-usage bar that warns before overflow — see where the money goes, live ### Archive & search - Zero import: install and your whole history is there, organized by project - Millisecond full-text search; can't recall the keyword? Ask in natural language and AI synthesizes the answer - AI-generated titles, tags, and summaries — stored outside the JSONL, of course ### Automation - "Summarize yesterday's sessions every morning at nine" — natural language becomes a scheduled task - Wakes the machine from sleep, runs, puts it back to sleep; authorize once, silent thereafter - Real hooks statistics: configured ≠ working — see which hooks actually ran in the last 7 days

Scheduled routines — natural-language instructions with cron schedules, run by the OS scheduler

### Always-on glances - Menu-bar subscription quota: Claude Code and Codex appear as vertically ordered provider sections, preserving the windows, groups, and reset countdowns each upstream actually supplies — visible with the main app closed - Desktop widgets: streaks, token pulse, 28-day heatmap, model mix - Cost estimation prices four token classes separately; unknown models are honestly labeled "unpriced" — never guessed

Desktop widgets — streaks, token pulse, 28-day heatmap, model mix, and cost estimation

### AI value-add (BYOAI) - Monet ships no model and charges nothing for AI — every capability runs on your own channels and quota, each toggleable, fully audited - Built-in MCP server: the Claude in your sessions can search your history, create scheduled tasks, and read Runner logs - 12 UI languages; name any other language and AI translates the entire interface on the spot ### Craft - Paper design language: warm, matte, ink-on-paper; a dark Ink theme included - No startup flash, ProMotion high refresh, virtual scrolling for huge sessions — details taken seriously, with a performance HUD (`Cmd+Shift+M`) keeping everything transparent

Permission checkup in the dark Ink theme — separate ledgers for the main app and background runner, every grant visible

## Install **Homebrew** (macOS): ```bash brew install --cask monet ``` Monet is available from the official Homebrew Cask repository, with no extra tap or `brew trust` step required. To migrate from the legacy `zenolab124/tap`: ```bash brew uninstall --cask monet brew untap zenolab124/tap brew install --cask monet ``` A regular uninstall removes only the app and Homebrew installation record; settings and data under `~/.monet/` remain intact. Or download from [Releases](../../releases): macOS `.dmg` / Windows installer. > macOS 11+ (Apple Silicon) gets the full experience; Windows covers the core features, with system-level integrations (widgets, menu bar, wake-from-sleep) being macOS-only. **First launch**: Monet is notarized by Apple — download and open, no extra steps. Users upgrading from early un-notarized builds are unaffected, and in-app updates continue as usual. **Works out of the box, zero import**: local Claude Code and Codex history is detected automatically and appears in one Archive, even when the corresponding CLI is not installed. Installing the Codex CLI adds interactive runtime features; inspect each engine under Settings → Engines. ## Build from Source ### Prerequisites - [Node.js](https://nodejs.org/) 22+ - [pnpm](https://pnpm.io/) 10+ - [Rust](https://rustup.rs/) 1.77+ - Xcode Command Line Tools — `xcode-select --install` ### Development ```bash git clone https://github.com/zenolab124/monet.git cd monet pnpm install pnpm tauri dev ``` ### Local Build ```bash pnpm tauri build ``` This builds the frontend, MCP binary, and main Tauri application. For official distribution packages, use the [Releases](../../releases) page. ## Data & Privacy | What | Where | Access | |------|-------|--------| | Claude Code sessions | `~/.claude/projects/` | **Read-only** | | Codex sessions | `$CODEX_HOME/sessions/` and `$CODEX_HOME/archived_sessions/` (default `~/.codex/`) | **Read-only; runtime actions use the official CLI App Server when available** | | Monet metadata (titles, tags, routines) | `~/.monet/` | Read-write | | MCP registration | `~/.claude/settings.json` | Adds `monet` entry under `mcpServers` | Local session features work offline, with zero telemetry and no Monet account system. Subscription quota contacts only the official Claude/Codex services and reuses the official clients' existing sign-in state; Monet never proactively refreshes or writes OAuth credentials. Credential discipline: API tokens never enter command-line arguments, and temporary files are deleted right after use. ## FAQ > Something broken? Skip the docs — hand this to your AI: `Read https://raw.githubusercontent.com/zenolab124/monet/main/llms.txt and help me troubleshoot Monet`. It can self-diagnose common issues and, if it's a real bug, file a report with diagnostics attached (no GitHub account needed). **Does Monet replace the Claude Code or Codex CLI?** No — it is a companion to those CLIs. The agents do the work; Monet gives you eyes and hands. Sessions started from a terminal or from Monet can be viewed together. **Is my session data safe?** Original session files are read-only by architecture, verifiable in the open source. Delete Monet and your Claude Code and Codex session data remain untouched. **Anything to configure after install?** Browsing existing history needs no extra setup. To create or continue sessions in Monet, the corresponding CLI only needs to be installed and signed in. Multi-channel support and AI enhancements are optional. **Gatekeeper warning on first launch?** There shouldn't be one — stable releases are notarized by Apple and open right away. If you are still blocked, you are running an early un-notarized build: grab the latest release, or grant a one-time exception under System Settings → Privacy & Security → **Open Anyway**. **Windows / Linux?** Windows is supported (core features complete, minus macOS system integrations); Linux has no near-term plans. **Which agents are supported?** Claude Code and Codex are supported today, with future engines able to join through the same contract. Developers can read the [Engine Adapter guide](ENGINE_ADAPTERS.md). ## Tech Stack - [Tauri 2](https://tauri.app/) — Rust backend + system WebView - [Vue 3](https://vuejs.org/) + TypeScript + Composition API - [UnoCSS](https://unocss.dev/) — atomic CSS (preset-wind4 + preset-icons) - [Shiki](https://shiki.style/) — syntax highlighting - [markdown-it](https://github.com/markdown-it/markdown-it) — Markdown rendering - [KaTeX](https://katex.org/) — LaTeX math rendering - [vue-i18n](https://vue-i18n.intlify.dev/) — internationalization - [@dnd-kit/vue](https://dndkit.com/) — drag and drop - [Swift WidgetKit](https://developer.apple.com/documentation/widgetkit) — macOS widgets ## Acknowledgments - Thanks to the [LINUX DO](https://linux.do/) community for the sharing culture and honest feedback - Thanks to every early user who filed issues and suggestions ## License [MIT](LICENSE)