Short answer
T3 Code and Orchestrator.inc (formerly Agent Orchestrator) are both free, open source apps for macOS, Windows and Linux, with phone apps, and both run the agent subscriptions you already pay for. T3 Code is a fast, chat-first control surface for a handful of agents, with easy remote access and broad Git host support. Orchestrator is a supervisor for many agents that follows each pull request through CI and review. Choose T3 Code if you drive a few agents yourself and want the simplest setup across phone, web and desktop. Choose Orchestrator if you run more sessions than you can watch and want CI and review feedback to reach each agent on its own.
At a glance
| Orchestrator.inc | T3 Code | |
|---|---|---|
| Best for | Supervising many agents through PR, CI and review | A fast GUI for a few agents on any device |
| License and price | Apache-2.0, free locally; optional Cloud plans from $20/month | MIT, free |
| Desktop | macOS, Windows x64, Linux | macOS, Windows, Linux, plus a CLI server with a local web app |
| Mobile | iOS and Android app with push notifications | iOS and Android app, plus a web app |
| Away from your network | Managed HTTPS tunnel; no built-in device relay yet | T3 Connect after signing in, with no port forwarding; LAN and Tailscale routes are learned |
| Coding agents | 38 harnesses, chosen per project and per task | Claude Code, Codex, Cursor, OpenCode and a few others |
| Isolation | Git worktree and branch per session; multi-repo workspaces | Optional worktree per task, cleaned up for you |
| Overview | Kanban board placed by session, PR, CI and review state | Thread sidebar |
| Git hosts | GitHub and GitLab, including self-managed GitLab | GitHub, GitLab, Bitbucket, Azure DevOps, Gitea, Forgejo and more |
| After the push | Failing checks, review comments and merge conflicts sent to the owning agent | PR list and review UI; "Work on PR" starts a chat on a PR |
| Planner agent | Project orchestrator that briefs and spawns workers | None |
The problem in practice
You have five agents working on one repository. Three have opened pull requests. One has a failing test, one has two review comments from a teammate, and one is waiting on a question. In a chat-first app each of those is a thread you have to open, read and nudge. With two agents that is pleasant. With five, you spend the afternoon deciding which thread to look at next.
How each tool handles it
Task: get a worker's PR from a failing test to merged.
In T3 Code
- Start a thread for the task, optionally in its own worktree, and chat with the agent until the change is ready.
- Open the PR and watch its checks. T3 Code shows pull requests and can turn on auto-merge while checks run.
- When a check fails or a reviewer comments, open the PR and choose "Work on PR" to start a chat about it, then tell the agent what to fix.
- Merge once checks pass.
In Orchestrator.inc
- Spawn a worker from the board or with
ao spawn. Once it pushes, Orchestrator discovers the PR. - The local daemon sees the failing check and sends its name, link and the end of the failed log to the worker that owns the PR.
- The worker pushes a fix. Review comments go to the same worker, and so do a reviewer agent's findings if you set one up.
- The card moves to Ready and you get a notification. You read the diff and merge; Orchestrator never merges on its own.
The difference is step three. In T3 Code you decide when to start the follow-up; in Orchestrator it arrives on its own.
Where T3 Code is a good fit
- You want remote access with very little setup. T3 Connect makes a machine reachable from your other devices after you sign in, and the app picks the fastest route.
- Your code lives on Bitbucket, Azure DevOps, Gitea or Forgejo. T3 Code covers more Git hosts than Orchestrator.
- You drive a few agents yourself. A chat-first app with a thread per task, and pull requests you can read on your phone.
Where Orchestrator fits better
- Workflows where CI and review feedback should reach the agent on its own. GitHub and GitLab, including self-managed GitLab, are polled with no webhooks to set up, and failing checks, requested changes and merge conflicts go to the owning session, de-duplicated.
- People who want a second model to review the first. 22 of the 38 harnesses can act as reviewer agents, so Codex can review a Claude Code PR.
- Larger plans. A project orchestrator agent breaks a plan into tasks, spawns and briefs workers, and follows their progress.
- People who run many sessions. Cards move across the board on their own, based on what is true of each session and its PR. Nothing is dragged by hand.
- Teams that use many agents. Each session picks from 38 harnesses, and each project sets its orchestrator, worker and reviewer agents.
Orchestrator reaches your desktop from the phone over a managed HTTPS tunnel and has no built-in device relay yet, so T3 Connect is simpler on that front. Orchestrator Cloud runs sessions in hosted sandboxes if you would rather not keep a machine on.
Pricing
Orchestrator is free and open source (Apache-2.0); you bring your own agent subscriptions. It uses the agent CLIs and sign-ins you already have, so usage is billed by your provider. Optional Orchestrator Cloud plans, for agents that keep working while your laptop is closed, are Starter at $20/month and Pro at $40/month, with Enterprise pricing on request. As of October 2026, T3 Code is free under the MIT license and also uses your own agent subscriptions.
Who should pick which
Pick T3 Code if
- you want the simplest remote setup across phone, web and desktop
- you use Bitbucket, Azure DevOps, Gitea or Forgejo
- you prefer a chat-first app and only use the agents it supports
Pick Orchestrator if
- you want CI failures, review comments and conflicts routed to the owning agent with no manual step
- you want a reviewer agent from another vendor on every PR
- you want a planning agent that spawns workers
- you run more than a few agents at once and want a board
Try it
Install Orchestrator, add a repository, and spawn your first two workers with the quickstart; then let one PR fail CI and watch the failure arrive in its session.
Frequently Asked Questions
Is T3 Code open source?
Yes. T3 Code is MIT licensed and free, as of October 2026. It reached its first stable release in October 2026. Orchestrator.inc is Apache-2.0.
Which agents does T3 Code support?
As of October 2026, T3 Code supports Claude Code, Codex, Cursor, OpenCode and a few other agents, using the subscriptions you are already signed in to. Orchestrator.inc supports 38 coding agent harnesses.
Does T3 Code have a mobile app?
Yes, on iOS and Android, plus a web app. Orchestrator.inc also has iOS and Android apps, with push notifications when a session needs you.
Does T3 Code fix failing CI or answer review comments automatically?
Not as documented in October 2026. It shows pull requests, supports auto-merge while checks run, and can start a chat on a PR when you ask. Orchestrator.inc sends failing checks and unresolved review comments to the owning agent automatically.
Can I use T3 Code and Orchestrator.inc on the same machine?
Yes. Both launch the agent CLIs you have installed, so you can try Orchestrator.inc next to T3 Code on one repository; the quickstart at docs.orchestrator.inc gets two workers running in a few minutes.