FAQ
Everything you need to know about how the radar works, scoring, badges, and API access.
- Open a submission using our Submit a Repo template. Submissions are reviewed and imported within a few days. The radar also runs automated 4-layer discovery (topics, keywords, trending activity, and free-text search) daily to discover emerging repositories automatically.
- • Processing pipeline: Runs 4 times daily (00:00, 06:00, 12:00, 18:00 UTC) to update stars, forks, issues, PRs, and re-compute Impact, Velocity, and Health scores.
• Discovery pipeline: Runs twice daily (02:00, 14:00 UTC) to discover newly published or trending AI projects.
• Website deployment: Automatically triggers after processing to publish the updated static dashboard to GitHub Pages. - Radar uses three-axis scoring: • Impact (40%): Percentile ranking of stars, forks, and watchers. • Velocity (35%): 7-day and 30-day growth rate and acceleration, not absolute star volume. • Health (25%): Commit recency, release cadence, issue turnaround, and maintainer count. If a project's star growth slows down relative to the prior week (e.g., from +100 stars/day to +20 stars/day), the acceleration drops, causing Velocity to decrease even while cumulative stars rise.
- Use the badge endpoint:
Replace[](https://radar.bharatdn.com/project/OWNER/REPO/)OWNER__REPOwith your repository path separated by two underscores (e.g.BerriAI__litellm). - Yes, all radar intelligence is published statically under
/api/:- Top projects:
/api/top.json - All repos:
/api/repos.json(and/api/repos.csv) - Category intelligence:
/api/categories.json - Anomaly alerts:
/api/anomalies.json - Breakout candidates:
/api/breakouts.json - Machine-readable dataset for LLMs:
/api/llms.txt
- Top projects:
- Scores are normalized 0–100 and bucketed into tiers:
- 90–100: Elite — Category-defining projects
- 75–89: Leading — Strong momentum & adoption
- 60–74: Rising — Healthy growth trajectory
- 40–59: Emerging — Early traction
- 0–39: Discovering — New or low activity
- Health (25% weight) combines:
- Commit recency: Days since last commit (lower = better)
- Release cadence: Frequency of tagged releases
- Issue turnaround: Median time to close issues
- Maintainer count: Number of distinct committers (more = resilient)
- Yes! Use the Compare Tool — enter up to 4 repositories to see Impact, Velocity, Health scores and historical trends side by side.
- Breakouts are projects with high Velocity + Health but lower Impact — they're accelerating fast but haven't hit mainstream visibility yet. The Breakouts page surfaces these rising stars early.
- Open an issue on GitHub with the Data Correction template, or submit a fresh entry via Submit a Repo — we'll re-process on the next pipeline run.
- Absolutely. All data is public via the API (JSON/CSV). We also publish
/api/llms.txtfollowing the llms.txt convention for AI agent consumption.
Still have questions?
Open a discussion or submit an issue — we're happy to help.