Apps

Ask us to fix a failed build

USES THE COMPANY'S AI CREDITS. Reads the build log, has our model write a minimal fix, stores it as a reversible overlay on the app's build (the repository itself is not changed — DELETE .../ai-fixes removes it) and redeploys with it; the settled result is { ok, summary }, where ok: false means there was nothing to offer. Only a failed deploy that is marked fixable; repeating it on the same failure spends credits again. Answers 202 { operationId, status: "running" } at once and runs the work in the background — it can take minutes. Poll GET /client/operations/{operationId} until it is done (the body is in result) or failed (error carries what the call would have refused with). id is a SERVICE id — take it from services[].id on GET /client/apps/{appId}, not the app id itself. Requires apps:deploy.

post/client/services/{id}/autofix

Request

  • Base URL: https://api.roar-ai.com
  • URL: https://api.roar-ai.com/client/services/{id}/autofix
  • Auth: HTTP bearer

Path parameters

idstring uuid required

Response

The fix has started — follow it on GET /client/operations/{id}

operationIdstring uuid required
status'running' required

Changes

No recorded changes to this operation across all 1 revision of this API.