Recently, since a lot of people have started using AI in their daily work, a problem has appeared: AI can make mistakes, and sometimes those mistakes slip through and get deployed to production. When that happens, many issues show up, and it becomes exhausting to find the problems and fix them. I thought about building something for code review. There are already many tools out there, but I wanted to focus on two things: making it very affordable and building it around custom rules that I define. Some might ask: how can free models give good results? The truth is, I put a small balance into Claude.ai and use their cheapest model to generate the prompt that I then send to weaker models. This way, I can still benefit from them and get an average result—much better than before and at a very low cost. That’s the idea behind Mouse Code. I’d love for you to try it and share your feedback, especially if you have suggestions for improvements.

Mouse Code

Your shortcut to smarter pull requests.

Your shortcut to smarter pull requests.