Launching today
To200
From production error to reviewable pull request
3 followers
From production error to reviewable pull request
3 followers
To200 captures production errors, groups repeated events, diagnoses likely root causes, writes branch-only fixes, runs available checks, and opens pull requests with the assessment and verification evidence attached. Start free with 1,000 credits and human approval, then move up to Pro, Scale, or High volume as usage grows.
Hi, I'm Albin — solo founder, and To200 exists because I was tired of being my own on-call firefighter.
The gap I kept hitting: monitoring tells you what broke, then leaves you the actual work. So I built the loop that comes after the alert — To200 captures the error, groups the repeats, diagnoses a likely root cause, writes a fix on a branch, runs your checks, and opens a PR with the evidence attached. You review and merge. It never touches main on its own.
The part I trust most I watched happen last night: it wrote a fix, ran it in a sandbox, caught that its own fix failed typecheck — and refused to ship it. That refusal is the product.
If the evidence isn't there, it stops and says what's missing.
Free is 1,000 credits on one repo, no card. GitHub required; Vercel and Supabase optional.
I'd genuinely value skeptical questions — especially about verification evidence, what failure classes it can't handle, and where the automation boundary should sit.