90
+340
// summary
Code-Review-Graph builds a structured map of the codebase using Tree-sitter, providing AI assistants with precise context to reduce Token consumption. The tool supports incremental updates, enabling rapid tracking of the impact scope of code changes during file edits or commits. It integrates with mainstream AI coding tools via the MCP protocol, significantly improving the efficiency and accuracy of code reviews.
// use cases
01
Automatically identify files affected by code changes through Blast-radius analysis, reducing Token waste from AI reading irrelevant code.
02
Support 19 programming languages and Jupyter Notebooks, completing updates for large codebases within 2 seconds via an incremental parsing mechanism.
03
Provide various MCP workflow templates including code review, architecture mapping, debugging, and onboarding to assist developers in smarter code reviews.