π Live Demo: https://repo-refine.vercel.app/
RepoRefine is a GitHub profile and repository audit tool that analyzes profiles, repository links, activity patterns, tech stacks, README quality, and project health to detect weaknesses and generate actionable improvements β making your GitHub presence recruiter-ready.
RepoRefine combines deterministic GitHub GraphQL checks with optional Groq-powered review summaries so scoring stays transparent while recommendations can still feel personalized.
- Profile completeness analysis
- Bio strength check
- Pinned repo validation
- Branding score
- Recruiter impression score
- Analyze a single GitHub repository from its URL
- Tech stack detection from languages, manifests, config files, topics, Docker, and CI signals
- README quality scoring with detected and missing sections
- License detection
- Maintenance activity check
- Open issue count detection
- CI/CD presence detection
- Test folder detection
- Documentation scoring
- Feature and update recommendations for the given repository
- Commit consistency detection
- Inactivity gap detection
- Commit message hygiene scoring
- Commit chronotype analysis
- Profile Health Score (0β100)
- Repository Health Score
- Documentation Score
- Consistency Score
- Recruiter Readiness Index
- Radar skill graph
- Health score meter
- Repository audit cards
- Red/Yellow/Green indicators
- Commit heatmap visualization
- Frontend: Next.js (App Router)
- Styling: Tailwind CSS
- Charts: Recharts
- API: GitHub GraphQL API and optional Groq API
- Auth: GitHub Personal Access Token
git clone https://github.com/Sushma-1706/RepoRefine.git
cd RepoRefinenpm installCopy the example file:
cp .env.example .env.localThen open .env.local and fill in your values.
- Go to github.com/settings/tokens
- Click Generate new token (classic)
- Give it a name (e.g.
RepoRefine) - Select these scopes:
read:userβ to fetch profile datarepoβ for repository analysis
- Click Generate token and copy it
- Paste it as the value of
GITHUB_TOKENin.env.local
β οΈ WithoutGITHUB_TOKEN, the app will silently fail with "Could not analyze profile."
- Go to console.groq.com
- Sign up or log in
- Create an API key and paste it as
GROQ_API_KEYin.env.local
Without
GROQ_API_KEY, AI review commentary will be replaced with a fallback message.
npm run devApp runs at:
http://localhost:3000RepoRefine uses the GitHub GraphQL API, which enforces rate limits on all requests.
| Auth status | Rate limit |
|---|---|
| Unauthenticated | 60 requests / hour |
| Authenticated (with PAT) | 5,000 points / hour |
RepoRefine makes multiple GraphQL queries per audit. Always use a GitHub Personal Access Token (PAT) to avoid hitting limits quickly.
- Audit results fail to load or appear incomplete
- You see a
403error or a message likeAPI rate limit exceeded - The app stops responding mid-analysis
1. Add your GitHub PAT to .env.local
GITHUB_TOKEN=your_github_pat_hereGenerate a token at github.com/settings/tokens. For read-only auditing, the following scopes are enough:
read:userrepo(for private repo analysis, optional)
2. Check your current rate limit status
Visit this URL while logged in to GitHub:
https://api.github.com/rate_limit
Look for "graphql" β "remaining" to see how many points you have left, and "reset" for when it refills.
3. Wait for the reset
Rate limits reset every hour. If you've hit the limit, wait until the reset time shown in the API response above.
- Avoid auditing the same profile repeatedly in a short window
- Use a dedicated PAT for RepoRefine rather than a shared one
- If you're running RepoRefine locally for development/testing, keep audits spaced out
RepoRefine uses GitHub GraphQL queries to fetch:
- Profile metadata
- Repository data
- Commit history
- Issues & pull requests
- Then it applies deterministic audit logic to:
- Detect weaknesses
- Calculate scores
- Generate structured improvement suggestions
- Identify the tech stack and README improvements for a pasted GitHub repository URL
If GROQ_API_KEY is configured, RepoRefine can also generate persona-based review commentary for profile audits.
- PDF export mode
- Before vs After improvement comparison
- Recruiter bulk candidate mode
- Open-source scoring benchmark
MIT License