Manages and maintains GPU-accelerated semantic search indexes for local codebases within Claude Code.
This skill provides comprehensive administrative control over semantic search indexes, allowing users to initialize, update, and troubleshoot local vector databases for enhanced code understanding. It streamlines the process of keeping code embeddings current through granular file-level reindexing or full codebase refreshes, ensuring Claude has access to the most accurate behavioral search results. By integrating with self-hosted Docker services like Qdrant, it offers a high-performance solution for developers who need deep, context-aware search capabilities across their private projects.
Key Features
01Granular file-level reindexing for immediate search updates
02Real-time collection status and metadata monitoring
03Multi-collection management for handling various project directories
040 GitHub stars
05Automated codebase indexing with GPU acceleration support
06Integrated troubleshooting for local service connectivity
Use Cases
01Initializing a semantic search index for a new local repository to enable behavioral search
02Updating specific file embeddings after major logic changes or refactors
03Diagnosing and resolving connectivity issues with local Qdrant or embedding services
