Thank you for helping reduce Claude Code costs for developers worldwide!
git clone https://github.com/Horilla/claudectx.git
cd claudectx
npm install
npm run build
npm testnpm run dev # Build in watch mode
npm test # Run tests
npm run lint # ESLint
npm run format # Prettiersrc/
├── commands/ # One file per CLI command (analyze, optimize, watch, mcp, compress, report)
├── analyzer/ # Context analysis (tokenizer, context-parser, waste-detector, cost-calculator)
├── optimizer/ # Auto-fix modules (claudemd-optimizer, ignorefile-generator, cache-advisor)
├── mcp-server/ # MCP server with smart_read, diff_read, index_query tools
├── dashboard/ # Live ink TUI dashboard components
├── compressor/ # Session memory compression
├── reporter/ # Usage reporting
└── shared/ # Types, constants, models, logger, config
We use asciinema + agg (NOT vhs — vhs is a Go tool, not available via npm):
brew install asciinema agg
asciinema rec --title "claudectx demo" demo.cast
agg demo.cast docs/demo.gif --theme monokai --speed 1.5- Fork the repo and create a feature branch:
git checkout -b feature/your-feature - Write tests for new functionality
- Run
npm testandnpm run lint— both must pass - Submit a PR with a clear description of the change and motivation
- Reference the relevant Jira issue (CTX-XX) if applicable
feat: add smart_read MCP tool
fix: correct token count for Unicode strings
docs: update README with v0.3.0 dashboard screenshot
test: add waste-detector edge case tests
chore: update dependencies
Use the bug report template.
If claudectx saved you money, share your numbers in Discussions and join the Hall of Fame!