An open-source bloatware cleaner powered by an LLM

Hi everyone, 🙂

I’m a junior developer and wanted to share a project I’ve been working on, hoping to get some feedback and ideas from this amazing community.

Project Name: Grayhound

• Why Grayhound?

I noticed that many bloatware cleaners rely on static blocklists, which often miss region-specific or newly popular grayware.

As a personal learning project, I created a tool that updates its own definitions by analyzing discussions from IT communities. This allows it to detect software that might only be considered "bloatware" in specific countries like the US, Korea, China, etc.

• What does it do?

-💡 AI-Driven: Uses Retrieval-Augmented Generation (RAG) with LLMs to build a dynamic, transparent threat database from tech forums.

-📖 Open & Transparent: 100% open-source (MIT license). It explains why a program is labeled as bloatware.

-✅ User in Control: Scans your PC, but every suggestion is reviewable before you take any action.

-🖥️ Platform: Currently works for Windows 11 only (Tauri frontend + Python backend).

• I'd love to hear your thoughts:

-Is this community-driven detection approach useful?

-Are there similar tools I should learn from?

-Any ideas on handling edge cases or making it more robust?

If you'd like to test it out or contribute, the code is open and PRs/issues are always welcome.

GitHub: https://github.com/RenardtheWolfdog/Grayhound/

Note: This tool requires admin rights for some actions. Please back up your data and review all detected items before cleaning.

Thank you for reading!