Archon is an open-source workflow engine for making AI coding work deterministic and repeatable. It lets teams encode development processes as YAML workflows that combine AI reasoning with fixed steps for planning, implementation, testing, review, approvals, Git operations, and pull-request creation.

Each workflow defines the sequence, validation gates, and artifacts while AI agents provide intelligence only where it is needed. Archon can isolate runs in separate Git worktrees, execute multiple jobs in parallel, mix shell or test nodes with AI nodes, and expose the same workflows through interfaces such as the CLI, web UI, Slack, Telegram, or GitHub.

The project is aimed at teams that want to move beyond one-off coding-agent conversations and standardize how automated development work is performed. By committing workflows alongside the codebase, organizations can review, version, and reuse their agent-driven engineering process much like CI/CD configuration.