Stop babysitting agents. Start merging real work.
Run a fleet of coding agents while keeping branches, reviews, and CI failures manageable.
Use the agents you already trust.
Tell the orchestrator what you need
Every project gets a main agent that runs the fleet for you. Describe the outcome, it plans the work, spawns workers into their own worktrees, keeps them moving, and escalates only what needs a human.
Agent OrchestratorThe whole fleet, at a glance
Every session flows across one board, working, needs you, in review, ready to merge, with its agent, branch, and PR state on the card. You see what needs attention and ignore the rest.
CI fails. The right agent fixes it.
AO watches every PR it opens. Failed checks and review comments route back to the session that owns the branch, with the context to fix them, until the PR is approved.
Use the agent you already trust
25 harnesses supported, with per-project agent choice. Claude Code, Codex, Cursor, OpenCode, Aider, Goose, and more, AO keeps it running the way you already run it while the tools underneath evolve.
Claude CodeYour fleet goes where you go
Pair the AO mobile app with your desktop over LAN or Tailscale. Watch every session, open a terminal, and get notified when an agent needs you - while execution and code stay on your machine.
Agents
192.168.88.3
4
working
1
need you
1
mergeable
See it in action
Watch AO run a fleet of agents end to end on a single repo, from task to merged PR.
In the wild
Real feedback from builders using AO.
I came across AO while using it during a hackathon. It was simple enough for me to get used to it within a couple hours. And I am glad I did. I had never handled more than one agent before. It handled context across workers it spawned and it could even communicate with other orchestrators in the workspace. Thinking of the agents' work as separate git worktrees also kept my mind at peace, because at the end, the decision is with me, whether I want to merge the branch or not.
AO helped me in removing the friction of using multiple harnesses at the same time. I can just work in peace without ever worrying about running grok, omp, or opencode CLI separately. I get the best of all worlds!
Before AO, I would ship at most 2–3 PRs a day. Now I consistently ship 5+ PRs every day at work.
AO automatically gets the right agent to address CI failures and review comments. My agents are much more autonomous now, and with the orchestrator + kanban, I’m able to manage more and more of them.