What is MemNexus?
MemNexus gives AI assistants persistent memory that works across every platform.
The Problem
AI assistants have amnesia. Every conversation starts from scratch — no memory of your preferences, your projects, or what you discussed yesterday.
The memory that does exist is siloed. ChatGPT Memory only works inside ChatGPT. Claude Projects only work inside Claude. If you use Claude, ChatGPT, Cursor, and Gemini, your context is fragmented across four platforms with no way to connect them.
This means you spend time re-explaining the same things, losing valuable context, and watching your AI assistants make the same mistakes repeatedly.
The Solution
MemNexus is a universal memory layer for AI assistants — what one AI learns, every AI can recall.
MemNexus stores memories in a graph database, makes them searchable via semantic search, and works with any AI platform through the CLI, SDK, or Model Context Protocol (MCP).
With MemNexus, your AI assistants can:
- Remember context across conversations and sessions
- Search past knowledge using meaning, not just keywords
- Learn your preferences by detecting behavioral patterns over time
- Share memory across platforms — what Claude learns, Cursor can recall
- Build knowledge graphs from facts, entities, and relationships
How It Works
You interact with MemNexus through three interfaces:
| Interface | Best For |
|---|---|
CLI (mx) | Developers managing memories from the terminal |
SDK (@memnexus-ai/sdk) | Applications integrating memory programmatically |
| MCP | AI assistants (Claude, Cursor) accessing memory natively |
Behind the scenes, MemNexus stores everything in a Neo4j graph database with OpenAI embeddings for semantic search. This means you can search by meaning ("what was that deployment issue last week?") and traverse relationships between memories, topics, and entities.
Key Capabilities
Memory Management — Store, search, update, and organize episodic memories with rich metadata, topics, and temporal context.
Semantic Search — Find relevant memories by meaning using vector embeddings, keyword matching, or hybrid search that combines both.
Knowledge Graphs — Store structured facts as subject-predicate-object triples. Build an explicit knowledge base your AI agents can reason over.
Pattern Detection — Automatically detect recurring topics, behavioral patterns, and preference trends from your memory history.
Conversation Tracking — Group related memories into conversations with automatic session detection and temporal queries.
GraphRAG — Advanced graph-based retrieval augmented generation that traverses relationships between memories, topics, and communities.
Who Is It For?
- Developers who want their AI coding assistants to remember project context
- Power users who work with multiple AI platforms and want unified memory
- Teams building AI-powered applications that need persistent context
- AI agent builders who need their agents to learn and remember
Next Steps
Ready to try it? Head to the Quick Start to create your first memory in under 5 minutes.