I wanted a quick way to evaluate repos before adding them as dependencies. ChatGPT can do this, but I needed consistent JSON output for automation, something API-callable for my agents, and deterministic results (no hallucinated star counts).

So I built RepoCrunch. Point it at a GitHub URL and get structured analysis: tech stack, dependencies, architecture, health metrics, security indicators.

What I found analyzing 8 popular frameworks:

- Next.js is 13% Rust. Turbopack is a big chunk. GitHub just says "JavaScript." - Flask has 3 open issues on 71K stars. The Pallets team runs the tightest ship in open source. - Express is still 100% JavaScript. No TypeScript migration. Sometimes "if it works, don't fix it" is the right call. - Go stdlib is 5.4% Assembly. Makes sense for a runtime, but you'd never know from GitHub. - 100% GitHub Actions across every framework. Travis CI didn't show up once.

Sample output (FastAPI):

  {"summary": {"stars": 94912, "license": "MIT"}, "tech_stack": {"framework": "Starlette", "key_deps": ["starlette", "pydantic"]}, "health": {"open_issues": 168, "contributors": 886, "commit_frequency": "daily"}}

Try it:

  pip install repocrunch
  repocrunch analyze fastapi/fastapi -p

Also has a built-in MCP server (repocrunch mcp) and REST API (repocrunch serve).

GitHub: https://github.com/kimwwk/repocrunch

Would love feedback - what metrics would make this more useful for your workflow?