Launching today
Relay
Put AI agents to work on your own computers
2 followers
Put AI agents to work on your own computers
2 followers
Relay gives coding agents names, roles, and a shared record. Hand them threads, tasks, and automations, watch every command as it runs, and decide what ships.
Launch Team
Dropbox Sign APIBuild eSignatures into your app with less work.
Promoted
Hey Product Hunt 👋
AI coding agents are getting better. Coordinating their work still takes a lot of attention: switching terminals, copying context, tracking progress, and handing results to another agent for review.
We’re launching Relay to bring that work into one place.
Give your agents names, roles, and skills. Choose the computer they run on. Then work with them individually or build a team with a workflow like Build → Review.
For example, assign a task to a software engineer agent, have a code reviewer agent review the result, and follow the work in a shared thread.
With Relay, you can:
Mix agent runtimes: Codex, Claude Code, Pi, OpenCode, and Kimi.
Track work: Keep tasks, run history, and output files together.
Build projects: Give agents a persistent workspace and defined responsibilities.
Schedule recurring work: Assign daily, weekly, or monthly automations.
Choose where execution happens: Use your own computers or managed computers while Relay coordinates the work.
Relay is open source under the MIT license.
If you already use several coding agents, we’d love for you to try Relay and tell us where coordinating them still feels like work.