Most AI dev tools have the same assumption baked in: your code gets sent somewhere. For a lot of teams, that’s not acceptable. So I built this mainly for myself: Noorlytics — a CLI that analyzes your codebase using AI, but runs locally by default.

What it actually does

• Flags technical debt patterns
• Analyzes dependencies & licenses
• Suggests refactors
• Generates unit test stubs Nothing groundbreaking individually, but bundled into a simple CLI you can run locally.

Why local-first

I didn’t want to push internal code to third-party APIs, deal with compliance/security reviews, or wonder how my code is being stored or used. So this runs with local LLMs (via Ollama). Cloud (OpenAI) is optional if you explicitly enable it.

Example usage

noor analyze
noor suggest
noor add-tests

Honest state

It’s early. I’m trying to figure out if this is actually useful beyond my own workflows. Would really appreciate feedback:

• Is this something you’d use?
• Where does it break down?
• What would make it actually valuable in your setup?

Links

• Landing page noorlytics.se
• Download: download zip and try it out

If you’ve avoided AI tools because of code privacy, this might be interesting. If not, I’m curious why.