Launched this week
Replicas
Run Claude Code and Codex in the cloud
344 followers
Run Claude Code and Codex in the cloud
344 followers
Run background coding agents from anywhere. Spawn Claude Code or Codex in a VM with code and tooling ready to go. Hand off tasks from Slack, Linear, or GitHub Replicas runs Claude Code, Codex, or any coding agent in the cloud. Agents run in isolated VMs with real dev environments, and you can bring your own subscriptions and API keys. Trigger tasks from Slack, Linear, or GitHub and come back to a PR ready to review.
Free Options
Launch Team / Built With
what happens when the agent gets stuck, goes in the wrong direction, or produces a PR that's technically complete but architecturally wrong. is there a way to see what the agent did and course correct mid-task or is it strictly a fire-and-forget model where you only see the output at the end
Refined Zapier
We’ve been using it and loving it! We kick off the cloud agents via Linear and Slack and it’s been seamless. The warming is key, we can run the app to test and full integration tests before it commits.
Replicas
@zlwaterfield Thanks Zach!
Been using Replicas for months and love it. This has let me kick off multiple changes and bug fixes in parallel without worrying about managing or coordinating test environments.
Oh, so excited to see you guys launching!
I’ve been using replicas for all my PR reviews. They have a great set of automations and my agent auto resolves.
Not to meant really frictionless integrations with Linear and Slack. I do most a lot of coding tasks from Slack now.
Running Claude Code and Codex in isolated VMs sounds powerful. How does startup time compare to local development environments?
spawning claude code in a VM from anywhere is something i didnt know i needed until just now. been running everything locally and the context switch when im not at my desk is painful. does it keep the session alive or does it reset each time
"Come back to a PR ready to review" is the big promise, but where exactly does it stop? If the agent hits a dependency it can't resolve — a type that doesn't exist, a build that fails repeatedly — does it leave a branch and ping you, or just stop silently?