Documentation

Everything you need to run greatmemory — from a first memory on your laptop to an authenticated server in the cloud, wired into your agents.

Getting started

Local quickstart

Install gmem, start the server, store and search your first memory — all on your own machine.

Server quickstart

Auth, CORS, SQLite vs Postgres, and Docker Compose for a server other machines can reach.

Agents

MCP server

Both transports, the five tools with parameters and examples, and space conventions.

Claude Code

Persistent memory for Claude Code with one command.

Codex CLI

Register gmem mcp in ~/.codex/config.toml.

Hermes

stdio MCP configuration for Hermes.

OpenClaw

stdio MCP configuration for OpenClaw.

Custom agents (REST)

curl, Python, and TypeScript clients for the memory loop, plus OpenAPI client generation.

Deployment

AWS

EC2 + Docker Compose, or ECS Fargate with RDS Postgres + pgvector.

Azure

Container Apps with Azure Files or PostgreSQL Flexible Server; Container Instances as the simple path.

Google Cloud

Cloud Run with Cloud SQL + pgvector, or a GCE VM with a stateful disk.

Reference

CLI reference

Every gmem command and flag, client environment variables, exit codes.

Architecture & integration

The layered design, the write and read paths, and patterns for integrating greatmemory into your systems.

API keys & air-gapped use

Creating, rotating, and verifying API keys — and running with no internet access at all.