Is that AI project still alive?
Stars tell you a project was popular once. They don't tell you it's maintained now. We measure the health of every tracked open-source AI project daily — straight from GitHub — so you know what's thriving and what's quietly abandoned before you build on it.
Checked · repositories · refreshed daily
How the health score works
The question everyone asks before adopting an open-source tool: is this project alive, or will I be on my own in six months? Nobody answers it clearly. Star counts say nothing — a repository with 40,000 stars can be abandoned for two years. That's the classic trap. So we measure four things that actually predict whether a project will still be there for you, refreshed every day from the GitHub API.
40 ptsActivity
When was code last pushed? This is the decisive signal. A project pushed within the last week scores full points; one untouched for over a year scores zero, whatever its star count.
25 ptsMomentum
How many commits landed in the last four weeks — and is that accelerating or slowing compared to the four weeks before? A trend arrow shows the direction.
20 ptsCommunity
How many people contribute? A single-maintainer project is fragile — the classic "bus factor" of one. Dozens of contributors means the project survives any one person walking away.
15 ptsMaintenance
Is the open-issue pile being managed? We weigh issues against popularity: 300 open issues on an 80,000-star project is normal; on a 500-star project, it's a project drowning.
What we don't do
No cosmetic scores. An archived repository is flagged as archived, full stop. A project with zero commits in the last four weeks and no push for over 30 days has its verdict capped, even with hundreds of historical contributors. One condition alone is not enough: 36 projects today have no commits in four weeks yet still read "Healthy", because they were pushed to recently. When GitHub doesn't return a data point, the component scores zero rather than being left out — a limitation we are working on, and one you should know about when reading a low score. That's the difference between a score that reassures and a score that informs.