stevepridemore/graph-memory

Overview

A personal knowledge graph for Claude that persists across sessions, devices, and tools. It uses Neo4j and semantic embeddings to store and retrieve information, secured via OAuth, and is single-user by default but designed for multi-tenant use.

Best for

Best for
Developers building persistent memory for Claude-based applications

Use cases

• Maintain long-term context for Claude conversations across sessions
• Store and query structured knowledge from multiple devices
• Build a persistent memory layer for AI assistants

Notes

A personal knowledge graph for Claude that persists across sessions, devices, and tools. It uses Neo4j and semantic embeddings to store and retrieve information, secured via OAuth, and is single-user by default but designed for multi-tenant use.

1 stars on GitHub. Last updated 2026-05-17. Licensed MIT.

Use cases

• Maintain long-term context for Claude conversations across sessions
• Store and query structured knowledge from multiple devices
• Build a persistent memory layer for AI assistants

Pros

• Survives across sessions and devices for consistent context
• Uses Neo4j and embeddings for flexible, semantic retrieval
• OAuth security with multi-tenant-ready architecture

Cons

• Single-user by default, requiring configuration for multi-user
• Requires Neo4j setup and management
• Limited community adoption with only 1 star

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.