Houseofmvps/codesight

Universal AI context generator. Saves thousands of tokens per conversation in Claude Code, Cursor, Copilot, Codex, and more.

What it solves

AI assistants often waste thousands of tokens by manually exploring a codebase to understand its structure, routes, and schema before they can answer a question. codesight reduces this overhead by pre-compiling a codebase into a compact, AI-optimized context map and knowledge base, significantly cutting token usage and improving response accuracy.

How it works

The tool uses AST (Abstract Syntax Tree) precision for TypeScript projects and battle-tested regex detection for other supported languages. It scans the project root to identify routes, schema models, UI components, and dependency graphs without requiring API keys or external LLM calls.

It provides three layers of context:

  1. Base Scan: Creates a compiled map of the project (e.g., CODESIGHT.md).
  2. Wiki Knowledge Base: Generates a persistent, targeted set of markdown articles (e.g., auth.md, database.md) so the AI can read only the relevant section rather than the full map.
  3. Knowledge Mode: Maps non-code documentation, such as ADRs, meeting notes, and Obsidian vaults, into a KNOWLEDGE.md primer.

Who it’s for

Developers using AI coding assistants like Claude Code, Cursor, GitHub Copilot, Windsurf, Cline, or Aider who want to reduce token costs and accelerate the AI's understanding of their project.

Highlights

  • Zero Dependencies: Runs via a single npx call with no configuration or API keys required.
  • Broad Language Support: Works with TypeScript, JavaScript, Python, Go, Ruby, Elixir, Java, Kotlin, Rust, PHP, Dart, Swift, and C#.
  • High Token Efficiency: Benchmarks show up to 91x reduction in tokens compared to manual AI exploration.
  • MCP Integration: Includes 14 MCP tools, allowing assistants like Claude Code to programmatically access the project map and wiki.
  • Blast Radius Analysis: Can identify all affected files, routes, and models when a specific file is changed.