Cortex Documentation

Complete guide to building AI agents with persistent memory

Welcome to the Cortex documentation! This guide will help you understand, integrate, and master Cortex in your AI agent systems.

📚 Documentation Structure

🚀 Getting Started

Start here if you're new to Cortex:

• Introduction - Mission, vision, and why Cortex exists
• Installation - Set up Cortex in your project
• Five-Minute Quickstart - Build your first memory-enabled agent
• Core Concepts - Understand the fundamentals
• Configuration - Configure Cortex for your needs

✨ Core Features

Deep dives into Cortex capabilities:

• Memory Spaces - Isolated memory banks for agents
• Semantic Search - AI-powered memory retrieval
• User Profiles - Store and manage user context
• Context Chains - Hierarchical agent coordination
• A2A Communication - Agent-to-agent memory tracking
• Conversation History - Persist message threads
• Access Analytics - Usage patterns and insights
• Fact Extraction - Storage trade-offs: facts vs conversations
• MCP Server - Cross-application memory sharing
• Hive Mode - Multi-participant agent coordination
• Fact Integration - Automatic fact extraction in Memory API

📖 API Reference

Complete API documentation (organized by architectural layers):

Getting Started:

• Overview - API conventions and patterns

Core Memory System (Layers 1-3):

• Memory Operations - Layer 3 convenience + Layers 1-2 overview
• Conversation Operations - Layer 1a: ACID conversations

User & Coordination:

• User Operations - User profiles + GDPR cascade ✅
• Agent Management - Agent registry + cleanup ✅
• Context Operations - Workflow coordination ✅
• A2A Communication - Agent messaging patterns

Advanced Storage (Layer 1b-c):

• Immutable Store - Layer 1b: Shared versioned data
• Mutable Store - Layer 1c: Shared live data

Supporting APIs:

• Governance Policies - Retention and compliance
• Memory Space Operations - Hive/Collaboration Mode management ✅
• Facts Operations - Structured knowledge extraction and storage ✅
• Graph Operations - Graph database integration ✅

Reference:

• Types & Interfaces - TypeScript definitions
• Error Handling - Error codes and debugging

🏗️ Architecture

Understanding how Cortex works:

• System Overview - High-level architecture
• Data Models - Convex schemas and indexes
• Convex Integration - How Cortex uses Convex
• Vector Embeddings - Embedding strategy
• Search Strategy - Multi-strategy retrieval
• Context Chain Design - Context propagation
• Agent Registry - Optional registry architecture
• Performance - Scaling and optimization
• Security & Privacy - Data protection

🔗 Integrations

Available integrations:

• Vercel AI SDK - Complete Vercel AI SDK integration guide

🎓 Advanced Topics

Expert-level content:

Graph Capabilities:

• Graph-Lite Traversal - Built-in graph queries and relationship navigation
• Graph Database Integration - Complete integration guide for Neo4j, Memgraph
• Graph Database Selection - Compare graph DB options
• Graph Database Setup - ✨ NEW Quick setup guide (<5 minutes)

Storage Optimization:

• Facts vs Conversations - Storage strategy trade-offs and efficiency
• Dimension Strategies - Choosing vector dimensions and models

🎯 Quick Links

I want to...

• Start from scratch → Installation
• Understand concepts → Core Concepts
• See API examples → Memory Operations
• Integrate with Vercel AI SDK → Vercel AI SDK Integration

📖 Documentation Conventions

Code Examples

All code examples are:

• Copy-paste ready - Run them directly
• TypeScript - Fully typed
• Commented - Explain the "why"
• Real-world - Not toy examples

Callouts

We use special callouts for important information:

Note: General information or tips

Warning: Important caveats or gotchas

Danger: Critical issues that could cause data loss

Version Indicators

• v0.1.0+ - Feature available since this version
• deprecated - Will be removed in future version
• experimental - API may change

🆘 Getting Help

Can't find what you need?

1. Search Docs - Use the search bar (top right)
2. Check FAQ - Common questions answered in [FAQ
3. GitHub Discussions - Ask the community
4. Discord - Real-time help in #help channel
5. Email - Contact support@cortexmemory.dev

🐛 Found an Issue?

Documentation problems? Please help us improve:

1. Typo or minor fix - Submit a PR directly
2. Missing content - Open an issue describing what's needed
3. Confusing section - Tell us what's unclear
4. Wrong information - Please report immediately

See the GitHub repository for contribution guidelines.

🌟 What's New?

Recent documentation additions:

• 2025-10-31: 🎉 v0.7.0 - Graph Database Integration

  • Complete graph database support (Neo4j, Memgraph)
  • Real-time sync worker with reactive queries
  • syncToGraph option across all APIs
  • Multi-layer context enrichment (2-5x more context!)
  • Sophisticated orphan detection
  • New: Graph Operations API
  • New: Graph Database Setup Guide
  • Updated: All API docs with syncToGraph options

• 2025-10-30: Complete Facts layer integration into Memory API with automatic extraction, cascade delete, and enrichment

• 2025-10-28: Graph-Lite capabilities, Graph DB integration guides, Fact Extraction, MCP Server documentation

• 2025-10-23: Initial documentation structure created

See [Changelog for full history.

📄 License

All documentation is licensed under CC BY 4.0.

The Cortex code is licensed under Apache License 2.0.

---

Last Updated: 2025-10-31

---

Ready to get started? Head to Introduction →