bugstack started life as an internal tool for my other startup Fuelscout because my customers would encounter production bugs when I was away that would ruin their experience. When I went camping and came home to find out my internal tool had fixed an error and shipped it to production I knew it was bigger than Fuelscout, so my little tool became bugstack.
bugstack monitors your production errors 24/7. When a user hits a bug, bugstack captures the error, pulls context from your repo, generates a minimal surgical code fix, runs it through CI, and delivers it as a pull request. If your CI passes and the fix meets your confidence threshold, it can auto-merge — or you can review the PR yourself. The workflow is: Capture → Fix → Deploy.
Technically:
You install an SDK (npm install @bugstack/sdk) and init with an API key Errors get fingerprinted for smart deduplication (no duplicate PRs for the same bug) The AI analyzes root cause using your actual codebase as context, then generates minimal diffs — not giant rewrites Fixes are syntax-checked, scope-checked, and committed to a new branch with a detailed PR Each project has independent confidence thresholds and you choose manual review vs. auto-merge Dashboard tracks the full lifecycle: Pending → Building Context → Generating Fix → Testing → Fix Ready → Auto-Fixed
Supports JavaScript (Express, Next.js), Python (Flask, FastAPI, Django), Ruby (Rails), and more. 14-day free trial.
I'd love feedback from anyone who deals with on-call rotations or production error triage. What would make you trust an AI to fix your prod bugs?
Product Hunt (we launched today): https://www.producthunt.com/products/bugstack