Welcome to Build Agent
Learn AI Agent engineering by comparing two production codebases: Claude Code (TypeScript) and Codex (Rust).
๐ฏ What You'll Learnโ
By analyzing 500K+ lines of production code from Anthropic and OpenAI, you'll master:
- 15 reusable design patterns for building AI agents
- Architecture trade-offs between different approaches
- Production-grade practices from real-world projects
๐ Quick Startโ
1. Understand Whyโ
Why compare two projects? - Learn the value of comparative learning
2. See the Big Pictureโ
Architecture Overview (coming soon) - Compare the complete architectures
3. Run Codeโ
Check out our examples - Start with the minimal agent
๐ Learning Pathsโ
Fast Track (2 hours)
- Introduction articles
- Core agent loop
- Run examples
Deep Dive (10 hours)
- All core practices
- Multiple examples
- One case study
Production Ready (20+ hours)
- Technical deep dives
- Complete agent build
- Optimization techniques
๐ค Contributingโ
Found an issue? Have a suggestion? Open an issue or contribute!
Ready to start? โ Why Compare Two Projects?