About Traycer
Traycer is an open source (MIT) desktop app where several coding agents, such as Claude Code, Codex and Cursor, share Tasks with specs, tickets and reviews.
What Traycer does
A Task holds one piece of work: its agents, files, and artifacts. Artifacts are the specs, tickets, stories, and reviews that outlast any single conversation.
Several agents can work in parallel, each in its own git worktree, so their changes do not collide.
Agents message each other across Tasks and machines. A remote host runs agents on another computer while you keep working from this one.
The iOS and Android apps follow and steer the same Tasks. They do not run a host: the agents keep running on your computers. You can share a Task with people you invite on every plan; sharing a Task with your whole team at once is the Sync plan.
Which coding agents it works with
You use your own subscriptions. The Free plan does not meter tokens or require inference credits for these agents, except optional Traycer Inference, which uses Traycer credits. The compatibility table is in the docs.
Claude Code
Codex
OpenCode
Traycer Inference
Cursor
Grok
Qwen Code
Kiro
Droid
Kimi
Copilot
Kilo Code
OpenRouter
Amp
Devin
Pi
Hermes Agent
Oh My Pi
Hugging Face
Reasonix
Antigravity
Command Code
Who it's for
Traycer is for developers who already use coding agents and want several of them on one Task, with the specs, tickets, and reviews in that same place. The Free plan is for individuals. Sync is for a team sharing a Task together.
How it's different
Claude Code or Codex alone is one coding agent in one session. Traycer runs several coding agents together on a shared Task, each in its own git worktree, with specs, tickets, and reviews, and agents that message each other across Tasks and machines.
Other products also give coding agents a workspace. Traycer's is the shared Task: the specs, tickets, and reviews the agents use together, hosts on your other machines, and messages between agents across Tasks and hosts. Side-by-side comparisons are on the site.
Pricing and license
Traction
11,000+ developers using Traycer Desktop, and 40,000+ tasks created in Traycer Desktop.
FAQ
What is Traycer?
Traycer is an open source (MIT) desktop app where several coding agents, such as Claude Code, Codex and Cursor, share Tasks with specs, tickets and reviews.
Is Traycer free?
Yes. The Free plan is for individuals and includes cloud sync, device switching, remote hosts, and the mobile app, with no token metering and no inference credits required for your own coding agents. Sync, for sharing a Task with your whole team, is $20 per user per month, or $16 per user per month billed yearly.
Is Traycer open source?
Yes. Traycer is open source under the MIT license, and the code is on GitHub.
Which coding agents does Traycer support?
Traycer supports Claude Code, Codex, OpenCode, Cursor, Grok, Qwen Code, Kiro, Droid, Kimi, Copilot, Kilo Code, OpenRouter, Amp, Devin, Pi, Hermes Agent, Oh My Pi, Hugging Face, Reasonix, Antigravity, and Command Code. You use your own subscriptions, and the Free plan does not meter tokens or require inference credits for them. Traycer also supports optional Traycer Inference, which uses Traycer credits.
Does Traycer work on Windows, Linux and mobile?
Yes. Traycer Desktop runs on macOS, Windows, and Linux. The iOS and Android apps follow and steer the same Tasks while the agents keep running on your computers.
How is Traycer different from Claude Code or Codex?
Claude Code or Codex alone is one coding agent in one session. Traycer runs several coding agents together on a shared Task, each in its own git worktree, with specs, tickets, and reviews, and agents that message each other across Tasks and machines. Comparisons with other agent workspaces are on the site.
Who makes Traycer?
Traycer is made by Traycer AI, Inc., in Walnut Creek, California. You can reach the team.
How do I get started?
Download Traycer for macOS, Windows, or Linux, sign in, and send a prompt to start a Task. The iOS and Android apps are on that page. The quickstart is in the docs.
