About RepoRadar
RepoRadar is an automated open-source intelligence site. Twice a day it collects the repositories trending on GitHub, records their star and fork counts, and publishes reports that stay online forever — so trends are visible over weeks and months, not just today.
Where the data comes from
- Trending lists are read from GitHub's public trending pages (daily, weekly, monthly).
- Repository facts — stars, forks, watchers, license, languages, topics, contributors, branches, tags, first commit and commit counts — come from the GitHub REST and GraphQL APIs.
- Activity and insight graphs use GitHub's statistics endpoints: per-day commit counts for the contribution heatmap, weekly additions and deletions for code frequency, the punch card for weekday and hour patterns, and participation for the maintainer versus community split.
- READMEs are rendered by GitHub and embedded directly, so you can read a project without leaving the page. Releases, recent issues, discussions, pull requests, Actions results, and the dependency graph (SBOM) come from their respective endpoints.
- Star history before we started tracking is reconstructed from public GitHub event archives and calibrated to each repository's current star count; from then on our own twice-daily snapshots extend the curve. GitHub no longer exposes per-star timestamps directly.
Signals and awards
GitHub does not expose profile achievements or trophies through any public API, so the awards shown on repository pages are computed by us from the data above. Each one maps to a stated rule — for example, "Breakout launch" means at least 3,000 stars within 120 days of creation, and "Outside contributions" means at least 70 percent of the last year's commits came from people other than the maintainer.
Jump detection
A star jump is flagged when a repository gains stars at more than three times its trailing seven-day daily average (with a minimum threshold to avoid noise). Jumps usually indicate a launch, a viral post, or a major release.
AI-generated sections
The "what it does", "use cases", and "getting started" sections are generated by a language model from each repository's README and metadata, then land in the site through reviewed pull requests. Everything else on a repository page is raw data from the GitHub API.
Update cadence
A morning run publishes the daily report and refreshes trending lists. An evening run refreshes star snapshots and detects jumps. Weekly digests are published on Sundays and monthly roundups on the first of each month.
Who built this
RepoRadar is built and maintained by Shashank Saxena, a software engineer building small, focused products. GitHub · LinkedIn · Portfolio