Open Source · Self-Hostable

Memory for
AI Agents

Graph-native persistent memory for autonomous agents. 90%+ token savings, sub-millisecond retrieval, and a portable open standard — self-hosted for free.

Open Dashboard →Self-Host Free ↓

A memory substrate, not a black box

Graph-native memory

Typed memories (episodic / semantic / procedural / working) linked by a real relation graph — causes, contradicts, supersedes, same_as and more.

Semantic + text retrieval

OpenAI-embedding vector search over Qdrant with a transparent text fallback, optional recency weighting, and namespace isolation.

Portable & open

Export/import the full graph as PMF (JSON, binary MessagePack, or AES-256-GCM encrypted) with Merkle + content-hash integrity. No lock-in.

Plugs into your agents

REST API, an MCP server (12 tools), TypeScript & Python SDKs, and a CLI — wire memory into Claude, your IDE, or any agent runtime.

Self-host in minutes — free

One docker compose up brings up the API, web UI, SurrealDB and Qdrant. Open source under Apache-2.0. Upgrade to Pro for federation and higher limits when you need them. See the benchmark →

See pricing →View on GitHub