RepoRank computes a composite credibility score from 5 weighted subscores: Maintenance (30%), Community (25%), Security (20%), Documentation (15%), and Adoption (10%). In total, 17 public GitHub data points feed into these subscores — including commit frequency, issue response time, contributor count, security policies, README completeness, and more. All data is sourced from public GitHub information.
Stars measure popularity, not health. A repository with 50,000 stars could be completely abandoned. RepoRank tries to answer a different question: is this project actively maintained and likely to be reliable? A repo with 500 stars and weekly commits is scored higher than one with 50k stars and no activity in two years.
When a repository has zero human reviews, RepoRank generates an AI-powered analysis. This is a starting point to give you some insight even for new or obscure repos — not a substitute for real-world experience. AI reviews are always labeled clearly so you know they're automated.
One review per user per repository. Users can vote a review helpful or unhelpful. Reviews with more helpful votes appear higher in the list. Authors can also mark a review as outdated if the repository has changed significantly since the review was written.
Scores are computed on-demand when a repository is first looked up and then cached. If you revisit a repo, you get the cached score. AI reviews refresh after 7 days if the repo is re-queried. Your human reviews remain until you update or remove them.
No. RepoRank is an independent project. We use the GitHub public API to fetch repository data, but we're not endorsed or supported by GitHub in any way.
Use this markdown snippet, replacing
Badge colors reflect the score: cyan ≥ 70, amber 40–69, red < 40, gray means no score yet.
owner and repo with your project details:[](https://www.reporank.online/github/owner/repo)
Badge colors reflect the score: cyan ≥ 70, amber 40–69, red < 40, gray means no score yet.
The best way to improve your score is to encourage people who've actually used your project to leave honest human reviews. You can also raise your score organically by adding a SECURITY.md file, a CONTRIBUTING.md, CI/CD pipelines, test coverage, and releasing regularly. Those factors directly influence the subscores.
When you sign in with GitHub, we store your username and avatar to display on your reviews. We do not store your GitHub access token after the session ends. We do not access private repositories.
Leave honest, thoughtful reviews of repositories you've used and trust. Share RepoRank links on projects you believe in. If you want to get involved further, check out our CONTRIBUTING.md on GitHub.