CodeGraph

About

CodeGraph solves the problem of AI coding assistants consuming excessive context by repeatedly searching and reading code during development. Instead of inefficiently grepping files, CodeGraph indexes a project once, allowing AI assistants like Claude Code, Cursor, and Windsurf to perform precise, token-efficient queries for identifiers, method signatures, and project summaries. This significantly reduces the context needed for code navigation, enabling AI to focus more on problem-solving and less on information retrieval, ultimately saving 50-80% of context window capacity for code searches.

Key Features

• Generates project summaries including auto-detected entry points and language breakdowns
• 0 GitHub stars
• Supports incremental updates for single files after changes, maintaining index freshness
• Smart Extraction using Tree-sitter for accurate code parsing and indexing of identifiers
• Instant access to method and class signatures without reading entire files
• Enables querying across multiple linked projects for comprehensive code navigation

Use Cases

• Reduce AI context window usage when performing code searches and navigation tasks
• Provide AI coding assistants with precise, line-number accurate definitions of functions, classes, and variables
• Quickly generate project overviews and summaries for AI to understand codebase structure and purpose