Run a fleet, not a task.
Every run gets its own git worktree. Point ten agents at ten problems at once — no collisions, no stepping on each other, no merge mess to untangle.
The command layer for your AI coding agents
You were the bottleneck.
One developer. A fleet of agents. Define your coding workflows once, then dispatch them from your terminal, Slack, or a GitHub comment — each running isolated, in parallel. The work scales without you at the desk.
$
curl -fsSL https://archon.diy/install | bash The keyboard was never the job. Shipping was. Stop being the one hand on one task — direct the work, and step up a level.
Ship from your phone.
Dispatch work from your terminal, Slack, Telegram, a GitHub comment, or the web. Kick it off, walk away, come back to finished pull requests.
Your best patterns, packaged.
Capture review, planning, fixes, and feature work as YAML workflows — multi-step DAGs with loops, gates, and conditions. Write the play once. Run it forever.
Bring your own model.
Claude Code, Codex, and more behind one command layer. Switch providers per workflow or per node. Archon orchestrates — you stay in control.
Define. Dispatch. Ship.
01
Define
Write the workflow once in YAML — or use the bundled defaults.
02
Dispatch
Fire it from anywhere. Each run spins up its own isolated worktree.
03
Ship
Agents do the work in parallel. You review the pull requests.
A day with Archon
Fire three agents from Slack on your phone. Walk the dog.
Come back to three pull requests.
Isolated worktrees mean they never collide. Multi-platform means you're never chained to the desk. That's the trade you always wanted.
Put the keyboard down.
Free, open source, and running on your machine in under a minute.