Enables high-performance semantic search for local codebases using Ollama and vector embeddings.
The grepai skill integrates advanced semantic search capabilities directly into Claude Code, allowing developers to find relevant code patterns using natural language rather than restricted keyword matching. By leveraging Ollama and the bge-m3 model, it creates a localized vector index for Java, Kotlin, JavaScript, and TypeScript projects, facilitating intelligent code discovery and automated context handoff. This skill manages the entire search lifecycle, including infrastructure verification, automated configuration via AI sub-agents, and real-time indexing of file changes to ensure your search results are always up to date.
主な機能
01Real-time file watching for continuous, background index updates
02One-click management for indexing, optimization, and CLI upgrades
03Automated project analysis and configuration using Opus-powered sub-agents
04Built-in infrastructure and health checks for search dependencies
0516 GitHub stars
06Local semantic search powered by Ollama and bge-m3 embeddings
ユースケース
01Providing high-quality semantic context to Claude during complex refactoring or feature development
02Locating specific implementation logic in large or legacy codebases using natural language queries
03Automating the setup of a high-performance vector search engine for local development environments
