ThinkTerm CLI
A Rust terminal with a multiplexer: agent sessions from several machines in one window, plus a skill for seeing adjacent panes
- 246 ★ on the repository
- Last commit
- GPL-3.0
Broad access
Why this level
- The CLI sends input and key presses into terminal panes, so an agent can run commands in any of them
- The mux server accepts client connections, including from a browser and other machines
- The agents panel reads processes and shell output to recognize them
Install
- Works in
Manual install
brew install --cask roversx/tap/thinktermFor macOS. Builds for other systems are in the repository releases.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- The CLI sends input and key presses into terminal panes, so an agent can run commands in any of them
- The mux server accepts client connections, including from a browser and other machines
- The agents panel reads processes and shell output to recognize them
This is third-party code. Review the repository files before installing.
What it does
ThinkTerm is an open-source terminal in Rust with a built-in multiplexer, built on WezTerm. The mux server holds shells, tools and coding agents, and clients (the desktop app, a TUI and the browser) connect to the same sessions. Local and remote machines come together in one workspace. The Agents panel lists recognized coding agents in one place with their work status and project, so you stop guessing which session is working, which waits for input and where a build finished. When the server updates, sessions, builds and agents keep running. Agent detection can be turned off in settings. The CLI exposes panes to automation: `thinkterm cli send-text` sends input, `send-key` presses keys, `get-text` reads output, and `wait` and `events` report when commands finish. The skill in the repository teaches Claude Code, Codex and Cursor to see and work with neighbouring panes (`npx skills add RoversX/ThinkTerm --skill thinkterm -g`).
Who it is for. Developers who run several coding agents across projects and servers and want their status next to the terminal.
Good fit when
- Several agents run in different projects at once and tabs turned into guesswork
- Sessions must survive a server update and a dropped connection
- An agent should read the neighbouring pane's output and send commands there
- You need access to the same sessions from a browser and another machine
- You want a panel with the git diff of the adjacent terminal
Not a fit when
- You only need a skill for the agent: the project replaces your terminal as a whole, installing it for one skill is pointless
- You need a licence other than GPL-3.0: the project is distributed under it
- You need a mobile client: iOS and Android versions are in development
Example request
Read the neighbouring terminal pane, wait for the build to finish there and summarize the errorsLimitations
The project replaces the terminal as a whole: the mux server, desktop client and web access build from Rust sources or install from a release. Licence GPL-3.0. Platforms: macOS, Linux, Windows, web and TUI, while iOS and Android are in development. ThinkTerm recognizes agents by heuristics, so an unknown tool may not appear in the panel. The agent skill installs with a separate `npx skills add` command.
How to disable. Remove the app (brew uninstall --cask thinkterm), stop the mux server and delete the skill from the agent's skills folder. The Agents panel can be disabled in settings.
FAQ
What is the multiplexer in ThinkTerm?
A server that holds shells, tools and agents while clients connect and disconnect. The workspace is reachable from the desktop, a TUI and the browser, and sessions survive a server update.
How does an agent work with panes?
Through the CLI: send-text sends input, send-key presses keys, get-text reads output, and wait and events report command completion. The skill teaches the agent to use it.
Which agents are recognized?
The Agents panel gathers recognized coding agents into one list with status and project. Detection can be switched off in settings.
How does it differ from WezTerm?
It is built on WezTerm and adds a multi-machine workspace, an agents panel, plugins and a browser client. By the authors' benchmarks output throughput is higher than WezTerm's.
How do I install it?
On macOS through Homebrew: brew install --cask roversx/tap/thinkterm. Prebuilt packages are in the repository releases, and cargo build --release builds from source.
Discussion
No questions or reviews yet.
Related
CLI toolsA plugin that teaches Claude Code and Codex to drive the agterm macOS terminal via agtermctl: sessions, panes, overlays, status
A native macOS terminal built on Ghostty with vertical tabs and notifications when an agent needs your input
A skill that teaches an agent to split Zentty windows into panes, read neighboring sessions, search history and notify you
Your machine's terminal, windows and desktop in a browser, plus an MCP server for agents to read and type into other sessions
Open Terminal gives AI agents a computer of their own with a REST API: shell, files, code, browser and desktop
A shared terminal for a human and an AI: one serial or SSH session, board output reading, GDB, file transfer and an MCP server
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| ThinkTerm (this card) | no | 246 | ||
| tmux-ide | yes | 567 | a tmux layer with a layout for agents; ThinkTerm is a terminal with its own multiplexer | |
| tmux-bridge-mcp | no | not required | 108 | an MCP bridge to tmux; ThinkTerm builds pane control into the terminal |
README in short
The README opens with download links, then explains the idea of machines fusing into one environment. It gives throughput benchmarks against WezTerm and kitty, multiplexer features, the agents panel, the automation CLI, plugins and docs.