Codex - AI-Powered GitHub Explorer

Slash repository onboarding time with AI-powered code insights in Chrome

Inspiration

We noticed that developers spend significant time trying to understand new codebases when onboarding to projects. With Chrome's new built-in AI capabilities through Gemini Nano, we saw an opportunity to create a tool that could help developers quickly grasp repository architecture and code functionality directly in their browser.

What it does

Codex is a Chrome extension that leverages Chrome's built-in AI (Gemini Nano) to help developers understand GitHub repositories. It provides:

  • Instant repository summaries including architecture and setup requirements
  • Code file explanations with context-aware analysis
  • Interactive AI chat for deeper understanding
  • All processing happens locally in Chrome - no external APIs needed

How we built it

  • Used Chrome Extensions framework with Manifest V3
  • Integrated Chrome's new Prompt API to access Gemini Nano
  • Implemented file analysis for both single files and repository-wide context
  • Created an intuitive UI that matches Chrome's design language
  • Built privacy-first architecture keeping all processing local

Challenges we ran into

  • Working with experimental Chrome AI features that are still in development
  • Managing the complex requirements for Gemini Nano (storage, GPU, etc.)
  • Designing an intuitive UI that works seamlessly with GitHub's interface
  • Handling large codebases while maintaining performance
  • Balancing between detailed analysis and quick response times

Accomplishments that we're proud of

  • Created one of the first extensions using Chrome's built-in AI capabilities
  • Developed a privacy-focused solution that processes everything locally
  • Built an intuitive interface that feels native to Chrome
  • Implemented context-aware code analysis
  • Achieved significant reduction in repository onboarding time

What we learned

  • How to work with Chrome's new AI APIs
  • Best practices for Chrome extension development
  • Techniques for processing and analyzing code repositories
  • Importance of user experience in developer tools
  • Balancing feature scope with performance

What's next for Codex

  • Support for more programming languages and frameworks
  • Enhanced repository analysis with dependency tracking
  • Integration with more GitHub features
  • Team collaboration features
  • Performance optimizations for larger codebases
  • Support for private repositories

Built With

Share this project:

Updates