Vec Memory

Provides graph-based semantic memory with vector search using SQLite vec0 and Ollama embeddings for MCP-based storage and querying.

typescript

0

GitHub Stars

typescript

Language

6 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "dlasky-mcp-memory-vec": {
      "command": "npx",
      "args": [
        "vec-memory-mcp"
      ],
      "env": {
        "OLLAMA_MODEL": "nomic-embed-text",
        "MEMORY_DB_PATH": "./memory.db",
        "OLLAMA_BASE_URL": "http://localhost:11434"
      }
    }
  }
}

Vec Memory MCP Server provides graph-based semantic memory with vector search, backed by SQLite vec0 and Ollama embeddings. It exposes a standard MCP interface for storing memories, traversing relationships, and performing semantic queries, with flexible transport options to fit streaming or HTTP-based client needs.

How to use

You connect to the Vec Memory MCP Server using your MCP client and perform memory operations, relationship operations, and semantic searches through the MCP API. Start by configuring the server in your client so it uses the same MCP endpoint and transports as your environment. You can work with memories, build graphs of relationships between memories, and run semantic queries to retrieve contextually relevant results.

How to install

Prerequisites you must have before running the server.

Node.js 18+
Ollama must be installed separately

Installation from npm (recommended) adds the server to your MCP client configuration. Include the following in your MCP client setup to automatically download and start the server when needed.

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

Alternatively, you can install from source and build the project locally.

npm install
npm run build

Available tools

Memory operations

Store, retrieve, update, delete memories, and perform semantic searches across memories.

Relationship operations

Create and query relationships between memories, including traversal to connected memories.