chunkhound

chunkhound

A codebase search tool with local-first prioritization.

MCPDevOpen source
Type
MCP
Transport
stdio
Open source
Yes
GitHub Stars
★ 1.3k
Source
mcp-github

Overview

ChunkHound is a local-first codebase search tool that extracts code architecture, patterns, and organizational knowledge. It supports multiple programming languages and integrates into various IDEs and AI assistants via the MCP protocol. Ideal for large monorepos, security-sensitive codebases, and multi-language projects. Users can search using natural language queries or regular expressions without uploading code to the cloud.

Capabilities

  • Semantic code chunking algorithm
  • Multi-hop semantic search
  • Natural language queries
  • Regular expression search
  • Support for 32 languages
  • Real-time indexing

Use cases

Code search within large monoreposLocal search in security-sensitive codebasesUnified search across multi-language projectsCode exploration in offline/isolated environments

Setup

Requires: Python 3.10+uv 包管理器API Key(可选)
curl -LsSf https://astral.sh/uv/install.sh | sh; uv tool install chunkhound

This information was compiled by AI from public sources and may contain inaccuracies — please refer to the source.

FAQ

Which programming languages does ChunkHound support?

Supports 32 languages including Python, JavaScript, and TypeScript.

Do I need an API Key?

API keys are not required for regular expression searches, but are needed for embeddings and LLM features.

Related skills