Introducing Labor0: a graph-native approach to AI engineering

We’ve launched Labor0, a new engineering product in The AI Platform family.

The AI Platform coordinates people and AI specialists through rooms, model routing, and reusable workflows. Labor0 starts from the same principle—agents should move independently when the work is clear, while people remain in control when judgment matters—but applies a different execution model to software engineering.

Labor0 turns broad engineering requests into bounded tasks connected by a dependency graph. It determines what is ready or blocked, then runs ready work through Codex, Claude Code, or OpenCode.

Hosted work runs in managed cloud environments and can be started from mobile. When Plan Mode needs a decision, Labor0 sends it through Web Push so the user can approve the plan, request changes, or stop the task without watching a terminal.

We’re also experimenting with a more direct integration between Labor0 and The AI Platform. Because The AI Platform includes rooms, specialists, model routing, workflows, mini apps, and other collaboration surfaces, we don’t want to reduce the integration to another chat endpoint. Designing it properly will require more time.

I’ve led Labor0’s development from the beginning, and this is only the first release.

Learn more about why we built it:

https://theaiplatform.app/blog/introducing-labor0/

Explore Labor0:

https://labor0.com/