概览
What is MCP Memory Server with Qdrant Persistence?
This MCP server provides a knowledge graph implementation with semantic search capabilities powered by Qdrant vector database. It stores entities and relations in a graph structure, persists data both in a local JSON file and as embeddings in Qdrant, and uses OpenAI for generating semantic embeddings. It is designed for developers building AI assistants or applications that need persistent, semantically searchable memory.
How to use MCP Memory Server with Qdrant Persistence?
Install dependencies with npm install, then build with npm run build. Set the required environment variables (OPENAI_API_KEY, QDRANT_URL, QDRANT_API_KEY, QDRANT_COLLECTION_NAME) and add the server configuration to your MCP settings file under mcpServers. The server exposes tools for creating and querying the knowledge graph and searching semantically.
Key features of MCP Memory Server with Qdrant Persistence
- Graph-based knowledge representation with entities and relations
- File-based persistence via a memory.json file
- Semantic search using Qdrant vector database
- OpenAI embeddings for semantic similarity
- HTTPS support with reverse proxy compatibility
- Docker support for easy deployment
Use cases of MCP Memory Server with Qdrant Persistence
- Give an AI assistant a persistent, searchable memory of entities and their relationships
- Enable semantic search across stored knowledge concepts and observations
- Maintain a dynamic knowledge graph that synchronizes between local file storage and a vector database
FAQ from MCP Memory Server with Qdrant Persistence
What runtime dependencies are required?
Node.js, an OpenAI API key, and a running Qdrant server with a configured collection are required.
Where does the data live?
Data is stored both in a local file (memory.json) holding the full graph structure and in Qdrant as semantic embeddings for similarity search.
What tools does the server expose?
The server provides tools for entity and relation management (create_entities, create_relations, add_observations, delete_entities, delete_observations, delete_relations, read_graph) and semantic search (search_similar).
How does synchronization between file storage and Qdrant work?
When entities or relations are modified, changes are written to memory.json, embeddings are generated via the OpenAI API, and the vectors are stored in Qdrant, keeping both systems consistent.
How is authentication and transport configured?
Authentication is handled via environment variables (OPENAI_API_KEY, QDRANT_API_KEY). The server supports HTTP and HTTPS connections to Qdrant and includes configuration for reverse proxies and custom SSL/TLS setups.
