Status: Coming soon based on beta feedback
This guide will cover strategies for using claude-cognitive on large codebases.
We plan to cover:
- Organizing 50k+ line projects into documentation
- Keyword strategies for complex systems
- Co-activation graphs for deep integrations
- Managing token budgets across many files
- Team coordination patterns
- Real-world examples (MirrorBot: 1M+ lines, 3200+ modules)
claude-cognitive was built on a 1M+ line production codebase.
Strategies documented here are battle-tested, not theoretical.
What's YOUR large codebase challenge?
- Monorepo with multiple projects?
- Microservices architecture?
- Deep nested dependencies?
- Multiple programming languages?
- Legacy code + new code?
Share your scenario: GitHub Discussions
We'll write this guide based on actual beta user needs.
For now, see:
- Getting Started - Works for any size project
- CUSTOMIZATION.md - Keyword mapping via
.claude/keywords.json