Self-Hosted Deployment

Deploy Engram Memory on your own infrastructure

Self-Hosted Deployment

Engram can be self-hosted for full data sovereignty. Everything runs locally — embedding, vector storage, and the recall engine — with no external API calls and no data leaving your network.

Deployment Options

Docker

The recommended deployment method. A single container bundles Qdrant (vector database), FastEmbed (local embedding model), and the MCP recall server.

Docker Setup Guide — Full instructions for deployment, configuration, data persistence, backups, and troubleshooting.

System Requirements

Requirement	Minimum
Docker	20.10+
RAM	4 GB
Disk	10 GB (model weights + vector storage)
CPU	2 cores recommended
Python	3.10+ (only if running MCP server outside Docker)

Quick Start

docker run -d \
  --name engram \
  -p 8585:8585 \
  -p 6333:6333 \
  -p 11435:11435 \
  -v engram-data:/data \
  engrammemory/engram-stack

Verify the stack is healthy:

curl http://localhost:8585/v1/connect

Then connect your MCP client:

claude mcp add engram-memory --transport http --url http://localhost:8585/mcp

Next Steps

Docker Setup — Full deployment and configuration guide
MCP Server — Connect AI agents via MCP
Platform vs Community — Compare self-hosted vs cloud deployment