# ChunkHound

A local-first codebase intelligence tool that gives coding agents and teams cited context from current code, git history and technical web research before they edit, review or debug.

ChunkHound turns a codebase, its git history and technical web research into cited context that agents and teams can use before they edit, review, debug or explain software. It supports research before editing by giving an agent grounded architecture context, relevant files, recent changes and external constraints; it turns branch diffs, commit ranges, tags or specific commits into cited briefs for reviewing large pull requests and drafting release notes; it turns symptoms and stack traces into likely code paths and recent changes for tracing bugs; and it connects local code to the external docs, APIs, issues and articles an implementation depends on.

Regex search works without any API key, while optional embeddings can come from VoyageAI, OpenAI or a local Ollama model, and its own reasoning can run through the Claude Code or Codex CLI with no separate key, or through the Anthropic, OpenAI or Grok APIs.

ChunkHound is a Python package that requires Python 3.10 or newer and the uv tool, and is released under the MIT license.

Website: https://chunkhound.ai