Corgi: terminal multiplexer for agents
Corgi is a terminal multiplexer for coding agents. It runs Claude Code, Codex, Pi, and OpenCode, and leaves them running when you detach.
Key features
- Starts Claude Code, Codex, Pi, or OpenCode from the agent list
- Shows working, needs-you, and done status, plus activity summaries
- Tracks context, tokens, and rate limits for those four agents
- Supports multiple accounts for Claude Code, Codex, and Pi
- Can resume a session after a crash or reboot
- Detach leaves processes running; run corgi again to reconnect
- Installs on macOS (Apple Silicon) or Linux (amd64)
- Controls the running session from any shell with list, split, send, and read
- Category
- AI Coding
- Who it's for
- Developers who run several coding agents in one terminal.
- Tags
- agentterminalcoding-agenterror-trackingfonts-icons
- Website
- corgi.computer
- Listed
- 2026-10-24
Frequently asked questions
What is Corgi used for?
Terminal multiplexer for coding agents. It tracks Claude Code, Codex, Pi, and OpenCode, and keeps panes running after you detach. Typical uses: starts Claude Code, Codex, Pi, or OpenCode from the agent list; shows working, needs-you, and done status, plus activity summaries; tracks context, tokens, and rate limits for those four agents.
Related in AI Coding
See all →- Claude Code Anthropic's terminal and IDE agent for editing a codebase.
- Cline Open-source coding agent that runs in the editor.
- Cursor AI code editor with agent mode, based on VS Code.
- OpenAI Codex OpenAI coding agent for the terminal, editor, desktop app, and Codex Web.
- OpenCode Open-source terminal coding agent.
- Orca Runs coding agents in parallel, each in its own git worktree. Free and open source, on macOS, Windows, and Linux.
