LanceDB
Provides a Model Context Protocol (MCP) server enabling LLM applications to store and retrieve memories using LanceDB, an embedded vector database.
概要
LanceDB is a Model Context Protocol (MCP) server that integrates LanceDB, an embedded vector database, to provide a semantic memory layer for Large Language Model (LLM) applications. It allows you to store text with vector embeddings for later retrieval, enabling LLMs to access external data sources and tools seamlessly. The server implements 'add-memory' and 'search-memories' tools, allowing for the addition of new memories and the retrieval of semantically similar memories based on a given query.
主な機能
- Uses sentence-transformers for embedding generation
- Configurable database path, collection name, and embedding provider
- Implements 'add-memory' tool for storing text with vector embeddings
- Implements 'search-memories' tool for retrieving semantically similar memories
- Supports semantic similarity search using a configurable threshold
- 5 GitHub stars
ユースケース
- Enhancing chat interfaces with memory retrieval capabilities
- Building AI-powered IDEs with access to contextual information
- Creating custom AI workflows with semantic memory support
Sponsored