Index and semantically search local code repositories with AST-aware chunking.
npx skills add https://github.com/cameronsjo/dev-toolkit --skill chunkhound-cameronsjo
Please help me install this Agent Skill. Skill: chunkhound Source: https://github.com/cameronsjo/dev-toolkit/tree/main/skills/chunkhound Command: npx skills add https://github.com/cameronsjo/dev-toolkit --skill chunkhound-cameronsjo
Local-first semantic code search enables you to understand and reason about large codebases by indexing and querying code semantically, eliminating the need for manual grep or guesswork.
Index your codebase with chunkhound index and then perform semantic searches with chunkhound search.
High-intent search queries and answers about installing and using this skill.
You can perform semantic code search by indexing your codebase locally and querying it with natural language. This approach uses AST-aware chunking to enable precise searches across thousands of files in a monorepo without relying on manual grep.
Yes, local code search supports fully on-prem and offline operation. By utilizing optional Docker or Ollama configurations, you can index and query sensitive codebase data without any external network calls.
Semantic code search integrates directly via an MCP server to expose search and research capabilities to LLMs. This allows Claude and other MCP-enabled setups to perform multi-hop analysis and structured output generation over your codebase.
The best way to onboard to unfamiliar poly-language codebases is using local semantic search to answer cross-file inquiries and perform architectural analysis. It eliminates guesswork by indexing multiple programming languages simultaneously for comprehensive queries.
Setting up local semantic code search requires Python 3.10–3.13, an MCP-enabled environment, and a project root containing a .chunkhound.json configuration file. Optional Docker or Ollama setups are available for offline operation.
AST-aware chunking for code search provides precise semantic context by understanding code structure, unlike exact text matching. It enables fast searches for exact matches alongside deep research analysis for understanding complex cross-file logic.
vespaai-playgroundDefine Vespa application packages, deploy to Docker or Cloud, and query via Python APIs.
0xRayAISearch and analyze code implementations across multiple repositories.
KuassSearch indexed GitHub repositories using natural language queries.
mishahaninTranslate natural language queries into structural AST patterns for code search.
