Hi HN, I built this to answer a simple question: can you fingerprint a developer's style from their actual source code?
HowYouCode scans your public GitHub repos and analyzes real code patterns — naming conventions, function sizes, comment density, nesting depth, error handling, and branch complexity. No GitHub stats or metadata, just your actual code.
The result is a personality card with an archetype (like "The Clean Architect" or "The Pragmatic Hacker"), scores, and comparisons against the average.
No login required. Just enter any GitHub username and wait ~30 seconds.
Fun profiles to try:
- torvalds (Linux kernel style)
- antirez (Redis creator)
- sindresorhus (npm legend)
- ThePrimeagen
Tech: Vanilla JS frontend, Express backend, 5 GitHub PATs for rate limiting, server-side response cache. Runs on a single €6/mo VPS.
The scoring is opinionated — I'd love feedback on what metrics actually matter for "code quality" vs just "code style".
Source analysis, not star counting. That's the idea.
marcelglaeser•2h ago
HowYouCode scans your public GitHub repos and analyzes real code patterns — naming conventions, function sizes, comment density, nesting depth, error handling, and branch complexity. No GitHub stats or metadata, just your actual code.
The result is a personality card with an archetype (like "The Clean Architect" or "The Pragmatic Hacker"), scores, and comparisons against the average.
No login required. Just enter any GitHub username and wait ~30 seconds.
Fun profiles to try: - torvalds (Linux kernel style) - antirez (Redis creator) - sindresorhus (npm legend) - ThePrimeagen
Tech: Vanilla JS frontend, Express backend, 5 GitHub PATs for rate limiting, server-side response cache. Runs on a single €6/mo VPS.
The scoring is opinionated — I'd love feedback on what metrics actually matter for "code quality" vs just "code style".
Source analysis, not star counting. That's the idea.