StructPR - Turn AI-sized diffs into fast, confident code reviews

by
StructPR is a GitHub App that reorganizes pull requests by business logic instead of file order. It groups changed files by context (auth, payments, migrations), briefs you on risk, flags missing changes like forgotten tests, and suggests reviewers; deterministic, ready in ~2 seconds. We also just launched Shield: free, private PR triage for open-source maintainers drowning in AI-generated PRs. Fifty pull requests, five worth reading. Free tier, no credit card.

Add a comment

Replies

Best
Hey Product Hunt 👋 I'm Allan, a solo founder. I started building StructPR earlier this year, as AI-authored PRs became more and more commonplace, and it became harder to review thousands of lines of code. The way we use PRs is changing, but unfortunately GitHub still shows them as one alphabetical wall of diffs. Reviewers quickly get overwhelmed: they scroll, skim, and approve on vibes. StructPR regroups the diff by what it means (authentication, payments, migrations, etc.), briefs you on risk, and catches missing pieces like forgotten tests. It's deterministic and pattern-based, so results land in about 2 seconds, not after an LLM round-trip. Alongside this we just opened Shield: free, private PR triage for open-source maintainers getting buried in machine-generated PRs. It scores incoming PRs on structural signals (not "AI detection"), keeps verdicts private to you, and never auto-closes anything. Fifty PRs come in; you focus on the five worth reading. We have four design-partner slots open this quarter: structpr.dev/shield Free tier for one repo, no credit card. And since we're early, the people who show up today genuinely get to shape the roadmap; tell me how your team reviews AI-era PRs, and what would make you trust an automated triage verdict.

Finally tried this on a gnarly migration PR and the auth vs payments grouping actually made sense for once. The missing test flag caught a real oversight I'd have shipped, which alone justifies installing it.

The business-logic grouping is genuinely useful, saved me from scrolling through 30 files to find what actually mattered. Shield sounds like a lifesaver for maintainers too.

Honestly the grouping by business logic is a nice touch, way more useful than skimming a flat file list. The risk brief caught a missing test on my last PR that I would have missed otherwise.

How does it decide what counts as "business logic" vs just file structure for projects that don't have obvious folders like auth or payments scattered across a monorepo?