I built tech-debt-visualizer to find the worst offenders in a vibe-coded codebase. In projects where people hack quickly and iterate fast, code tends to bloat. Functions balloon, docstrings disappear, and complexity sneaks in everywhere.

Most tools either count lines or require a heavy setup, which misses the real hotspots. This works by utilizing both static and LLM-based approaches to determine the cleanliness of a codebase.

I'd love if you all gave it a try, and I am looking for any and all thoughts. You can try it out without even installing in one CLI command in your codebase right now: npx tech-debt-visualizer analyze .

(for the LLM part to work, you will also need an API key for Gemini, OpenAI, Openrouter, or OpenAI compatible)