Paperclip PR Reviewer
https://github.com/paperclipai/pr-reviewer- Category
- Developer Tools
- Rank
- No. 2010Tools index
- Pricing
- Open Source
- Type
- TOOL
- Builder
- paperclipai
- GitHub
- 14 stars
- Date
About
PR triage UI for reviewing pull requests with AI assistance.
What it does
Paperclip turns one repository’s pull request backlog into an ordered work queue. It syncs GitHub data, combines CI, conflicts, discussion, change size, contributor history, tests, issue links, documented reasoning, and freshness, then exposes results through a dashboard, HTTP endpoints, and a command line. Optional Anthropic analysis adds stored recommendations, risk labels, reasoning, and summaries.
Why it's ranked here
The tool has a clear operational opinion: review small, tested, conflict-free changes with healthy CI and useful context first. Its weighting logic is centralized and documented, while every result can include a signal breakdown. That makes its decisions inspectable rather than mysterious. The narrow repository target and reliance on external services keep it from being a general-purpose solution.
What's good
The priority model uses concrete maintenance signals and excludes bot comments from human discussion counts. It rewards test files, linked issues, and documented reasoning, while penalizing conflicts and weak contributor merge history. Local SQLite and Cloudflare D1 share one database interface. The command line supports syncing, filtering, detail views, and optional batch analysis. Configuration validation catches incomplete D1 credentials before work begins.
Tradeoffs
The GitHub owner and repository are fixed to Paperclip, so reuse elsewhere requires code changes. Contributor history can push authors down the queue, which may encode past acceptance patterns into future attention. Test detection relies on filename patterns, while issue and reasoning detection use text matching. Anthropic analysis only receives the pull request title and body in the shown code, not changed files or diffs. D1 batch writes run sequentially.
How to use it well
Use it when maintainers need a shared, explainable order for a busy Paperclip pull request queue. Sync GitHub data, filter out failing or conflicted work, inspect each signal breakdown, then reserve Anthropic analysis for the strongest candidates. Tune the centralized weights when team priorities change and update the documentation alongside them. It does not replace code inspection, CI, or a general multi-repository review platform.
Technical notes+
package.json defines a TypeScript build, Vitest scripts, an npm command-line binary, Hono, Octokit, Anthropic, SQLite, and Cloudflare dependencies. src/cli.ts implements sync, list, show, and review commands. src/scoring.ts centralizes weights, detection patterns, contributor calculations, and breakdown generation. src/config.ts validates SQLite or D1 configuration with Zod. src/db/client.ts selects a backend behind the interface in src/db/types.ts; src/db/sqlite.ts uses WAL and transactional batches, while src/db/d1.ts calls Cloudflare’s REST API and runs batch statements sequentially. src/web/app.ts mounts API and dashboard routes, src/web/server.ts adds a local sync endpoint, and src/llm/review.ts stores validated Anthropic output.
Observed
- License
- ISC
- Primary language
- TypeScript
- Packaging
- npm package with TypeScript build and command-line binary
- Interfaces
- Command line, Hono web dashboard, and JSON HTTP API
- Database support
- Local SQLite or Cloudflare D1
- Platform target
- Node.js locally and Cloudflare Workers for deployment
- External integrations
- GitHub through Octokit and optional Anthropic analysis
Read from README.md, package.json, src/cli.ts, src/config.ts, src/display.ts, src/scoring.ts, src/db/d1.ts, src/web/app.ts, src/db/types.ts, src/db/client.ts, src/db/sqlite.ts, src/web/server.ts, src/github/api.ts, src/llm/review.ts, src/web/routes.ts.
What it can do
Review pull requests with AI assistance
Pull request code changes → AI-generated code review comments and suggestions
Triage pull requests by priority
Multiple pull requests → Prioritized list of pull requests for review
Analyze code changes for potential issues
Code diffs and file changes → Issue detection report with recommendations
Generate automated review summaries
Pull request metadata and code changes → Structured review summary document
Filter and organize pull requests
Repository pull request data → Filtered and categorized pull request lists
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.