Latch
BuildingScores any public GitHub repo for AI-agent and newcomer readiness — with the exact fixes that raise the score.
Built by
Cansin YildizAbout Latch
Latch provides a readiness readout for any public GitHub repo to assess AI-agent and newcomer readiness. It returns a score per signal and lists the exact fixes that raise the score. The project is currently building and uses Next.js, React, TypeScript, Node.js, Tailwind CSS, and is hosted on Vercel. Latch reads a repo from top to bottom and reports a score per signal, using a six-category, 600-point scale.
Key features
- Readiness readout for any public GitHub repository
- Per-signal score with precise fixes that raise the score
- Six-category scoring system on a 600-point scale
- Each signal is scored 0–100
- Point at a repo to generate the readout
Tech stack
Frequently asked questions
What does Latch do?
Latch runs a readiness readout on any public GitHub repo and reports a per-signal score plus the fixes that raise it.
Who is it for?
It is for assessing AI-agent and newcomer readiness of a repository.
What tech stack does it use?
Next.js, React, TypeScript, Node.js, Tailwind CSS, and Vercel.
How is the scoring structured?
There are six signals on a 600-point scale, with each signal scoring 0–100.
