Trending on GitHub

This tool constructs a local, persistent structural map of your codebase using Tree-sitter to optimize how AI coding assistants process reviews. By tracking changes incrementally, it provides precise context through the Model Context Protocol (MCP), ensuring AI tools read only relevant code segments rather than scanning entire repositories. The project highlights benchmarked reductions in context size to lower token usage and improve review efficiency.

• Uses Tree-sitter for incremental structural mapping of codebases
• Delivers precise context to AI via MCP to reduce token waste
• Benchmarks show significant context reduction in large repos
• Supports CLI, local-first architecture, and GitHub Actions
• Optimizes AI review workflows by limiting scope to relevant changes

Jane Street has open-sourced Incremental, a library designed to handle incremental computations. This tool allows developers to build systems where outputs are automatically updated when inputs change, avoiding full recomputation. It is particularly useful for complex data processing pipelines and interactive applications where efficiency is critical.

• Enables automatic output updates when inputs change, reducing recomputation overhead.
• Ideal for complex data pipelines and interactive UIs requiring real-time responsiveness.
• Open-sourced by Jane Street, bringing institutional-grade OCaml tools to the public.
• Simplifies state management in applications with dynamic data dependencies.