qdrant-mcp

# Qdrant MCP Server A Model Context Protocol (MCP) server that provides semantic memory capabilities using Qdrant vector database with configurable embedding providers. ## Features - **Multiple Embedding Providers**: - OpenAI (text-embedding-3-small, text-embedding-3-large, text-embedding-ada-002) - Sentence Transformers (all-MiniLM-L6-v2, all-mpnet-base-v2, and more) - **Semantic Search**: Store and retrieve information using vector similarity - **Flexible Configuration**: Environment variables for all settings - **MCP Tools**: Store, find, delete, and list operations - **Metadata Support**: Attach custom metadata to stored content ## Installation ### Via uvx (Recommended for MCP) The server is designed to be lightweight by default. When using OpenAI embeddings: ```bash # For OpenAI embeddings (lightweight, no ML dependencies) uvx qdrant-mcp ``` For local embeddings with Sentence Transformers: ```bash # For local embeddings (includes torch and other ML libraries) uvx --with sentence-transformers qdrant-mcp ``` ### Via pip (Development) ```bash # Clone the repository git clone https://github.com/andrewlwn77/qdrant-mcp.git cd qdrant-mcp # Basic install (OpenAI embeddings only) pip install -e . # With local embeddings support pip install -e . sentence-transformers ``` ## Configuration The server can be configured using environment variables: ### Required Environment Variables - `EMBEDDING_PROVIDER`: Choose between `openai` or `sentence-transformers` - `EMBEDDING_MODEL`: Model name for the chosen provider - `OPENAI_API_KEY`: Required when using OpenAI embeddings ### Optional Environment Variables - `QDRANT_URL`: Qdrant server URL (default: `http://localhost:6333`) - `QDRANT_API_KEY`: Qdrant API key (optional) - `COLLECTION_NAME`: Qdrant collection name (default: `mcp_memory`) - `DEVICE`: Device for sentence transformers (default: auto-detect) - `DEFAULT_LIMIT`: Default search results limit (default: 10) - `SCORE_THRESHOLD`: Minimum similarity sc…