memorygraph-mcp-tai040502 0.12.5

Graph-based MCP memory server for AI coding agents with intelligent relationship tracking and multi-backend support

MemoryGraph MCP Server

Graph-based Model Context Protocol (MCP) server that gives AI coding agents persistent memory. Store patterns, track relationships, retrieve knowledge across sessions.

Installation

Docker

docker pull tai040502/memorygraph:latest

uv (pip)

uv tool install memorygraph-mcp-tai040502

MCP Configuration

Docker (stdio)

{
  "mcpServers": {
    "memorygraph": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "-v",
        "memory-data:/data",
        "tai040502/memorygraph"
      ]
    }
  }
}

uv

{
  "mcpServers": {
    "memorygraph": {
      "command": "uv",
      "args": [
        "tool",
        "run",
        "memorygraph-mcp-tai040502"
      ]
    }
  }
}

CLI Options

memorygraph                    # Core mode (default, 9 tools)
memorygraph --profile extended # Extended mode (12 tools)
memorygraph --backend neo4j   # Use Neo4j backend
memorygraph --version          # Show version

Backends

Backend	Type	Zero-Config
sqlite (default)	Embedded	✅
falkordblite	Embedded	✅
ladybugdb	Embedded	✅
falkordb	Client-server	❌
neo4j	Client-server	❌
memgraph	Client-server	❌
turso	Cloud	❌

Environment Variables

Variable	Description	Default
MEMORY_BACKEND	Backend type	sqlite
MEMORY_SQLITE_PATH	SQLite DB path	~/.memorygraph/memory.db
MEMORY_TOOL_PROFILE	Tool profile (core/extended)	core
MEMORY_LOG_LEVEL	Log level	INFO

License

MIT License