tmux-ide CLI
A CLI and terminal app that show which coding agents in your tmux sessions are working, waiting on you or done
- 558 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- The send command types text straight into another pane's prompt, and the screen it reads may hold secrets
- `tmux-ide integration install claude` writes hooks into `~/.claude/settings.json`, and `tmux-ide adopt` adds a status bar, a dock and keys to tmux
- A background daemon watches panes in every session and connects to remote machines over SSH, with a local control socket under ~/.tmux-ide
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/tmux-ideDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- The send command types text straight into another pane's prompt, and the screen it reads may hold secrets
- `tmux-ide integration install claude` writes hooks into `~/.claude/settings.json`, and `tmux-ide adopt` adds a status bar, a dock and keys to tmux
- A background daemon watches panes in every session and connects to remote machines over SSH, with a local control socket under ~/.tmux-ide
This is third-party code. Review the repository files before installing.
What it does
tmux-ide shows every coding agent running in your tmux sessions and the state of each: working, blocked, done or idle. tmux still owns every process and pane, so closing tmux-ide never stops an agent. The `tmux-ide app` command opens a full-screen interface with two surfaces. Home lists agents across local and SSH machines with search and status filters, and Terminals shows the current session with window tabs and pane headers. Status comes from two layers. An agent can report it itself through the `@agent_state` tmux option, and Claude Code gets ready hooks so new sessions report state exactly. When no report exists, tmux-ide resolves the agent from the process tree and reads the pane screen. When an agent blocks or finishes, you get tmux, macOS and terminal notifications with sound. The CLI is built for scripts and for agents themselves. `tmux-ide team --json` prints the state of all sessions, `tmux-ide send` types text into another pane's prompt, and `tmux-ide wait` blocks until a status or a line in the output. Scoped pane reads and sends are also available over MCP and an SDK. After a tmux crash one command rebuilds the windows and revives Claude Code, Codex, Cursor and opencode conversations.
Who it is for. For developers who keep several Claude Code, Codex and other agents in tmux panes, including on remote machines over SSH.
Good fit when
- Three Claude Code and two Codex agents run in tmux panes, and you want to see at once which one is blocked and waiting for your decision
- The server rebooted or tmux died, and you want the windows and agent conversations back without rebuilding them by hand
- A coordinator agent hands tasks to other agents with send and waits for them to finish with wait agent-status
- Some agents run on a remote machine over SSH, and you want it listed next to local ones
- Every branch needs its own git worktree and session, which the tmux-ide worktree create command sets up
Not a fit when
- You do not use tmux: tmux-ide works only on top of its sessions
- You need an orchestrator that assigns tasks between agents itself: tmux-ide shows and connects agents but does not run them
- You are on Windows without WSL: the app is built for macOS and glibc Linux
- A tmux server you already run is older than 3.7: tmux-ide never replaces a running server
Example request
Install tmux-ide, hook up Claude Code and show me with tmux-ide team --json which agents are working now and which one is waiting for my answerLimitations
- tmux-ide does not run agents itself. It shows and connects them, and coordination through `tmux-ide send` and `tmux-ide wait` is up to you or your lead agent.
- The app runs on macOS 26 or newer for ARM64, macOS 15 or newer for x64 and glibc Linux for ARM64 and x64. On Windows use WSL. The release bundles tmux 3.7c, but tmux servers you already run must be version 3.7 or newer.
- Exact Claude Code status requires installing the hooks. For agents without a verified resume story, such as Gemini, Aider and Copilot, the session id is self-reported.
- Version 2.9 of the app has no Files, Diff or Missions views.
How to disable. tmux-ide unadopt with a session name removes the dock and status bar, and the session stays plain tmux. The Claude Code hooks come out with tmux-ide integration uninstall claude. The package itself goes with npm uninstall -g tmux-ide or by deleting the folder where the installer put it.
FAQ
How do I watch several Claude Code and Codex agents in tmux?
Install tmux-ide and run `tmux-ide app`. On Home (key F1) you see every agent with its state and the Working and Needs attention filters, and F2 opens the current session. An agent that needs an answer is marked blocked, and Enter jumps to its pane.
How do I get exact Claude Code status in tmux-ide?
Run tmux-ide integration install claude. It adds hooks to ~/.claude/settings.json that mark the pane working, blocked, done or idle on each lifecycle event. It affects new Claude Code sessions, and tmux-ide integration uninstall claude removes the hooks.
How do I restore agent sessions after tmux crashes?
First preview the plan with tmux-ide restore --dry-run --json, then run tmux-ide restore --resume-agents. It rebuilds windows from the last snapshot and revives Claude Code, Codex, Cursor and opencode conversations whose session id is known.
How can an agent report its own status to tmux-ide?
Use tmux set-option -p @agent_state with a value like working:1730000000, where the number is a timestamp in seconds. Valid states are working, blocked, done and idle. A working or blocked stamp older than about ten minutes counts as stale, and tmux-ide falls back to reading the screen.
How do I install tmux-ide without sudo?
Run curl -fsSL https://tmux-ide.com/install.sh | sh. The installer brings its own Node.js, and both installs bundle tmux 3.7c, so neither sudo nor a system tmux is needed. If Node.js 20 or newer is already there, npm install -g tmux-ide works.
Discussion
No questions or reviews yet.
Related
CLI toolsA tmux-based TUI that runs Claude Code, Codex and other agents in one list: status, quick prompts, worktrees and diff review
A Go terminal multiplexer with tiled panes that shows the state of Claude Code, Codex and other agents and gathers their requests in one list
A terminal session manager for Claude Code, Codex, Gemini CLI, OpenCode and other agents: one list with statuses, git worktrees and control from a phone
A control plane that runs Claude Code, Codex and Gemini CLI agents in parallel with a shared task board and schedules
A terminal app for running several Claude Code, Codex, Gemini CLI and Aider sessions in parallel, each in its own tmux session and git worktree
botmux connects Lark and Feishu to Claude Code, Codex, Gemini CLI and OpenCode: each message starts its own agent session with live cards
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Repository stars | Main difference |
|---|---|---|---|
| tmux-ide (this entry) | yes | 558 | |
| Agent Deck | no | 1,180 | a terminal session manager for AI agents |
| Claude Squad | no | 8,694 | a terminal app that launches parallel agents itself |
| tmux-claude-hatch: Claude Code in a tmux popup | no | 403 | a tmux plugin that runs a Claude Code session in a popup for each |
README in short
The README starts with installation and a feature list, then covers agent teams and a key table. Separate sections cover Claude Code hooks, the optional window layout in workspace.yml and the architecture layers: tmux, a daemon per machine and the app. It ends with links to the docs and a manual written for agents.