orch

Run AI coding agents in parallel over a task graph, cap spend per provider, and give your client a live page instead of a status thread. One Go binary, local-first.

GitHub

Official source

Start locally with the CLI guide