Hey everyone! I recently revisited an old project after getting frustrated with my job, and I'm excited to see it evolving with the current AI tech. I've been working on this for about five months now and using a couple of Claude agents alongside Gemini to refine the responses. Things seem to be coming together, but I often run into problems where the AI makes incorrect assumptions about my setup.
I've been utilizing repomix for managing my main repos, but as my combined token count is nearly 950,000, it's tough for Claude to handle. Now that I'm in the testing phase, I'm considering tools that can interact with my VSC or run through my console to improve efficiency.
I've thought about having Gemini analyze my repos to create a detailed testing guide for Claude, but I keep hearing about other AI coding tools I haven't tried yet. Also, as a side note, I'd love any suggestions for checking the security of my software before I launch. For context, my project involves a desktop client and a webpage that communicate via WebSockets on my DigitalOcean servers. Any advice would be appreciated!
1 Answer
You can definitely use Gemini within Cursor. Personally, I find it to be much faster and user-friendly compared to Claude Code, especially for codebases over a million tokens. It's been my go-to lately! If you still find yourself needing an architecture summary, consider checking out DeepWiki. They offer impressive architecture and documentation generation. You do need to sign up for a $20/month plan for private repositories, so keep that in mind!
Is Claude Code directly available for Windows? I remember it being a bit tricky last time I looked.