Now in Public Beta

Long-term memory for AI agents

Help your AI applications remember user preferences, context, and relationships across conversations. One API call to store, one to retrieve. It's that simple.

Start Building Free View Documentation

app.ts

import { MagnetoClient } from '@magneteco/client';

// Initialize client
const memory = new MagnetoClient({ apiKey: 'your-api-key' });

// Store a memory
await memory.memorize({
  userId: 'user_123',
  content: 'User prefers dark mode'
});

// Retrieve relevant context
const context = await memory.retrieve({
  userId: 'user_123',
  query: 'user preferences'
});

Everything you need

Memory infrastructure for AI

Stop building memory systems from scratch. Magneteco handles extraction, storage, and retrieval so you can focus on your AI application.

Simple API: Two endpoints: memorize and retrieve. Store conversations, get back structured context. No complex setup required.
Vector Search: Powered by pgvector for lightning-fast semantic search. Find relevant memories based on meaning, not keywords.
Knowledge Graph: Neo4j-backed entity relationships. Track people, places, preferences, and how they connect.
Evolving Summaries: Category summaries that update as new information arrives. Always have the current state, not just history.
Fast Retrieval: Sub-second response times with intelligent caching. Get context when you need it without slowing down your app.
Multi-Tenant Secure: Complete data isolation between apps and users. SOC 2 compliant infrastructure with encryption at rest and in transit.