$ curl -fsSL https://metateam.ai/install.sh | sh && metateam

Metateam is a CLI tool that embeds Claude Code, Codex CLI, and Gemini CLI into its own process and unifies them into a single cohesive crew of agents with shared memory and a common communication system.

~/projects/myapp

Run One Command

Run metateam in your project directory. A terminal dashboard launches with your crew coordinator ready. Tell it what you need — fix a bug, write tests, refactor a module — and a team of AI agents coordinates the work. Switch between agent tabs to watch progress. Jump in to steer when you want. Review reports when they are done.

Your Agents Remember

Every session is captured automatically. When a new session starts, agents receive relevant context from past work and your curated Knowledge Base entries. Write a rule once, mark it important, and every future session starts knowing it. Summon a specialist weeks later and it picks up where it left off.

Your Team Coordinates

Agents message each other, write reports, and coordinate through the crew leader. You see their communication in real time from the dashboard. Steer any agent directly by switching to its tab. Review detailed results in the mail-board when the work is done.

How It Works

  1. Install Metateam and run metateam in your project.
  2. A dashboard launches with your crew coordinator. Tell it what you need.
  3. Agents work, sessions are captured, knowledge persists across every future session.

Get Started

Run the install command above, then launch metateam.

FAQ

Does it upload my source code?

Metateam does not read, index, or upload your repository. It uploads session transcripts — your conversation with the AI. Any code in the conversation is part of the transcript. Local code search stays on your machine. You control capture with .nometateam and .yesmetateam.

What AI clients are supported?

Claude Code with hooks, Gemini with hooks, and Codex with background monitor. Hooks give automatic context at session start.

Where is my data stored?

Session transcripts and Knowledge Base entries are stored in the Metateam API. Local artifacts like crew history, reports, and config stay on your machine.

Can I use it on multiple machines?

Yes. Sessions and Knowledge Base entries sync through the API, so agents on any machine share the same project memory. Cross-machine crew coordination is an evolving feature — check the docs for current capabilities.