Memory Qdrant MCP

An MCP (Model Context Protocol) server that provides memory management capabilities using Qdrant vector database for storing and retrieving project context, decisions, progress, and patterns.

Features

• Memory Management: Log and query project memories across different categories
• Vector Search: Semantic search through memory entries using embeddings
• Multiple Providers: Support for Gemini, Ollama, and FastEmbed embedding providers
• MCP Integration: Full MCP stdio server implementation
• REST API: Additional HTTP endpoints for direct access

Installation

Using npx (Recommended)

npx memory-qdrant-mcp

This will download and run the server automatically.

Manual Installation

npm install -g memory-qdrant-mcp
memory-qdrant-mcp

From Source

git clone <repository-url>
cd memory-qdrant-mcp
npm install
npm run build  # if needed
node server/index.js

Prerequisites

• Qdrant Database: The server requires a running Qdrant instance

  Option 1: Simple Docker run

  docker run -p 6333:6333 qdrant/qdrant

  Option 2: Docker Compose (Recommended for production) Create a docker-compose.yml file:

  services:
    qdrant:
      image: qdrant/qdrant:latest
      container_name: qdrant
      restart: unless-stopped
      ports:
        - "6333:6333"
        - "6334:6334"
      environment:
        QDRANT__SERVICE__CORS: "true"
      volumes:
        - qdrant_data:/qdrant/storage
  
  volumes:
    qdrant_data:
      driver: local

  Then run:

  docker-compose up -d

• Environment Variables: Copy .env.example to .env and configure:

  cp .env.example .env

  Embedding Provider (choose one):

  # Option 1: Google Gemini (default, fast, recommended)
  GEMINI_API_KEY=your_gemini_api_key_here
  
  # Option 2: Ollama (local, free, slower)
  OLLAMA_BASE_URL=http://localhost:11434
  
  # Option 3: OpenRouter (no embedding models available - use Gemini or Ollama above)

  Model Configuration (optional, defaults provided):

  EMBEDDING_MODEL=models/text-embedding-004     # Gemini default (use nomic-embed-text:v1.5 for Ollama)
  SUMMARIZER_MODEL=openai/gpt-oss-20b:free      # OpenRouter default (use gemini/gemini-2.0-flash-exp for Gemini, Any model in ollama is slow)
  DEFAULT_TOP_K_MEMORY_QUERY=3                  # Search result limit

  Required:

  QDRANT_URL=http://localhost:6333

MCP Configuration

For VSCode GitHub Copilot

Create or update the MCP settings file at:

• Windows: %APPDATA%\Code\User\globalStorage\github.copilot-chat\settings\mcp.json
• macOS: ~/Library/Application Support/Code/User/globalStorage/github.copilot-chat/settings/mcp.json
• Linux: ~/.config/Code/User/globalStorage/github.copilot-chat/settings/mcp.json

Add the following configuration:

{
  "mcpServers": {
    "memory-qdrant-mcp": {
      "command": "npx",
      "args": ["memory-qdrant-mcp"],
      "env": {
        "QDRANT_URL": "http://localhost:6333",
        "GEMINI_API_KEY": "your_gemini_api_key_here",
        "OPENROUTER_API_KEY": "your_openrouter_api_key_here",
        "OLLAMA_BASE_URL": "http://localhost:11434",
        "EMBEDDING_MODEL": "models/text-embedding-004",
        "DEFAULT_TOP_K_MEMORY_QUERY": "3",
        "SUMMARIZER_MODEL": "openai/gpt-oss-20b:free"
      }
    }
  }
}

For Roo

Add to your Roo MCP settings:

{
  "mcpServers": {
    "memory-qdrant-mcp": {
      "command": "npx",
      "args": ["memory-qdrant-mcp"]
    }
  }
}

Available Tools

log_memory

Log a memory entry to the vector database.

Parameters:

• project_name (string): Name of the project
• memory_type (string): Type of memory (productContext, activeContext, systemPatterns, decisionLog, progress)
• content (string): Content to log
• top_level_id (string, optional): Optional top level ID

query_memory

Query memory entries from the vector database.

Parameters:

• project_name (string): Name of the project
• query_text (string): Query text for semantic search
• memory_type (string, optional): Optional memory type filter
• top_k (number, optional): Number of results to return (default: 3)

log_decision

Log a decision entry.

Parameters:

• project_name (string): Name of the project
• decision_text (string): Decision text
• top_level_id (string, optional): Optional top level ID

log_progress

Log a progress entry.

Parameters:

• project_name (string): Name of the project
• progress_text (string): Progress text
• top_level_id (string, optional): Optional top level ID

summarize_text

Summarize the given text.

Parameters:

• text (string): Text to summarize

Publishing to npm

To publish your own version:

1. Update package.json with your information:

   • Change name to a unique package name
   • Update author, repository, homepage
   • Ensure version is appropriate

2. Login to npm:

   npm login

3. Publish:

   npm publish

4. Users can then install and run:

   npx your-package-name

Development

Project Structure

memory-qdrant-mcp/
├── server/
│   ├── index.js              # Main MCP server
│   ├── mcp_tools/            # MCP tool implementations
│   │   ├── memoryBankTools.js
│   │   ├── store.js
│   │   └── summarizer.js
│   ├── embeddings/           # Embedding providers
│   │   ├── providerBase.js
│   │   ├── geminiVertex.js
│   │   ├── ollama.js
│   │   └── fastEmbed.js
│   └── config.js             # Configuration
├── memory-bank/              # Project documentation
├── package.json
└── README.md

Adding New Tools

1. Implement the tool in server/mcp_tools/
2. Register it in server/index.js
3. Update this README

License

MIT