Quick Start (self-hosted)
Run Engram on your own hardware with Docker Compose. Up in five minutes.
Welcome to Engram
AI agent platform with persistent memory
What is Engram?
Engram is an AI agent platform built for people who want persistent, searchable memory and real operational control. It combines three components: OpenClaw (the agent runtime), memcp (the memory server), and claw-control (the management dashboard).
What makes it different?
Most AI tools treat memory as a feature bolted on after the fact. Engram was designed memory-first. memcp uses FSRS (a spaced-repetition algorithm) to weight memories by salience, and hybrid BM25+vector search to recall by meaning AND exact terms. Agents remember what matters, forget what doesn’t, and surface the right context at the right time.
How to get started
Self-hosted users can be running in under five minutes with Docker Compose. Hosted beta users get the same stack managed for them — no ops required. Choose your path below.
Join the Hosted Beta
Get managed Engram infrastructure — no Docker, no ops, just agents.
memcp reference
Learn how memcp stores, weights, and recalls agent memories.
Architecture
Understand how OpenClaw, memcp, and claw-control fit together.