Introduction
Welcome to the Memory Platform 2.0 — the production-grade, multi-dimensional memory backbone for AI Agents.
The Memory Platform gives your AI the ability to remember. Instead of starting every conversation from scratch, your AI can recall facts, preferences, and insights from previous interactions across any domain — from procurement and legal to personalized consumer assistants.
🚀 Platform 2.0 Capabilities
- Multi-Dimensional Navigation: Traverse memories across platform-defined axes (Users, Projects, Tenants, Cases) using a unified navigation context.
- Coordinate-Aware Document Intelligence: Link every derived fact back to its exact source in a document with 0.0-1.0 normalized coordinate mapping.
- High-Performance Visualization: Explore thousands of semantic relationships in real-time using our WebGL-powered Sigma.js graph explorer.
- Enterprise-Grade Isolation: Strict tenant segregation at the persistence level (JanusGraph & Qdrant) ensuring commercial-safe multi-tenancy.
- Protocol First: Built from the ground up on the Model Context Protocol (MCP) for seamless integration with LLM agents.
📦 The Ecosystem
The platform is composed of 11 specialized repositories working in harmony:
| Repository | Responsibility |
|---|---|
| memory-domain | Canonical data contracts and extensible Pydantic models. |
| memory-engine-service | The core brain service and API orchestrator. |
| memory-mcp-server | MCP-compliant interface for AI agents. |
| memory-store-janusgraph | The Knowledge Graph layer (JanusGraph). |
| memory-store-qdrant | The Semantic/Vector layer (Qdrant). |
| memory-store-documents | Coordinate-aware document parsing and source mapping. |
| memory-graph-visualizer | WebGL Admin Graph Explorer & Configuration Hub. |
| memory-curator-worker | Background LLM extraction and summarization worker. |
| memory-sdk-js | TypeScript/JavaScript SDK. |
| memory-sdk-python | Python SDK. |
| memory-infra | Infrastructure-as-Code and deployment orchestration. |
🛠️ Next Steps
- Quick Start: Get your first memory stored in 5 minutes.
- Core Concepts: Understand the mental model behind Memory Blocks and Dimensions.
- Architecture: Deep dive into the technical design.