FlowProof
AI that tests and repairs broken business workflows
Start new thread

FlowProof - AI Fullstack Engineer, now self-healing on Google Cloud Run

by•
Rebuilt serverless for the Cloud Run Hackathon: FlowProof now deploys and heals itself on Google Cloud Run. The same AI fullstack engineer that turns SOPs into continuously tested workflows — executing steps in a browser, detecting documentation drift, and recovering broken paths — now scales to zero and spins up on demand.

Add a comment

Replies

Best
Maker
📌
We built FlowProof because business workflows keep silently breaking when software changes under them — SOPs go stale, nobody notices until something expensive fails. So we made an AI that actually runs the workflow in a browser, catches the drift, repairs the path, and verifies the fix. For this launch we rebuilt the whole thing serverless on Google Cloud Run as part of the Cloud Run Hackathon: agents spin up on demand, scale to zero, and the repair loop is now self-healing infrastructure. Happy to talk architecture, agent design, or Cloud Run setup — ask us anything!
Google Cloud Run
Google Cloud Run Hackathon
How did you use Cloud Run for this launch?
Maker
We rebuilt FlowProof's execution layer on Google Cloud Run. Each workflow run spins up as a stateless container: browser-automation agents execute SOPs in sandboxed containers that scale to zero between runs, while a Cloud Run service hosts the API and dashboard. Scheduled health checks re-execute workflows, and when documentation drift is detected, a repair job runs in its own container and verifies the fix before proposing updates. This gives us per-run isolation, usage-based billing, and self-healing infrastructure with no idle servers — the whole product is Cloud Run-native end to end.