Appearance
GAIA orchestrates AI agents through the same ticket lifecycle your team already follows, state by state, with a human wherever you want one.
The agent reads the ticket, classifies it as bug or feature, and publishes a versioned handoff before any work starts.
Design decisions, an implementation plan, and checkable acceptance criteria, committed to the repo as the contract for the build.
Implemented in an isolated git worktree on its own branch; the acceptance checks run before the agent hands the ticket over.
Fresh eyes give an independent verdict on the diff, with the full run transcript available right on the ticket.
The merge request lands, the ticket closes, and the status is mirrored back to the customer's origin issue.
Run one on every machine your code runs on.
Isolated branches, no stepping on each other.
No orphaned runs, no client-side cleanup.
Workflow skills are versioned in your repo.
A ticket moves through explicit states. An agent qualifies, specs, codes; a human reviews. Ownership is your call, per state, and changing it is a config edit, not a re-org.
A conductor is a daemon on any machine you own. It registers itself, claims tickets it can work, creates a git worktree per ticket, and drives the agent to the next state.
A shared control plane, conductors on your machines, the agents you already use, and a workflow that lives in your repo.
The shared source of truth for projects, tickets, and runs. Your conductors talk to it over an API; the cockpit UI reads from it live.
A daemon on any machine you own. It registers itself, claims tickets it can work, creates a git worktree per ticket, and drives the agent.
The AI CLIs doing the work: Claude, Codex, Kimi, OpenCode, Pi. Each is an addon; swapping agents is a config line, not a migration.
The workflow itself, written as versioned skill files in the repo. What "qualify" or "review" means is code-reviewed, not tribal knowledge.