MemoryGraph gives your AI assistants persistent memory. Store solutions, track patterns, recall context across sessions. Zero config, local-first, privacy by default.

What is MemoryGraph?

MemoryGraph is an MCP (Model Context Protocol) server that provides persistent memory capabilities to AI assistants like Claude, GPT, and custom agents. It stores memories in a knowledge graph, preserving relationships between concepts.

Key Features

  • Persistent Memory - Your AI remembers across sessions
  • Knowledge Graph - Relationships between memories, not just a list
  • MCP Compatible - Works with Claude Code, Claude Desktop, and more
  • Local-First - SQLite by default, your data stays on your machine
  • Zero Config - Works out of the box with sensible defaults

Quick Start

Get up and running in 3 commands:

# Install MemoryGraph
pipx install memorygraphMCP

# Add to Claude Code CLI
claude mcp add memorygraph memorygraph

# Restart Claude Code - done!

See the Quick Start Guide for detailed instructions.

How It Works

MemoryGraph exposes MCP tools that your AI assistant can use to store and recall memories: 

# Store a memory
store_memory(
  type="solution",
  title="Redis timeout fix",
  content="Increased connection pool size to 20...",
  tags=["redis", "timeout", "production"]
)

# Recall memories
recall_memories(query="timeout solutions")

# Create relationships
create_relationship(
  from_memory_id="problem_123",
  to_memory_id="solution_456",
  relationship_type="SOLVES"
)

Next Steps