Quick Start

1. Install and start

Download the desktop app from the latest GitHub release and launch Agent Brain. Click Start Stack in the dashboard. Wait for all service indicators to turn green.

First launch can take 30-60 seconds while Docker pulls images and local models warm up.

2. Source install

git clone https://github.com/blackrock-engineering/agent-brain
cd agent-brain
cp .env.example .env
./scripts/setup.sh
docker compose -f docker-compose.yml -f docker-compose.dev.yml up -d

Verify the stack:

curl http://localhost:9090/health

3. Connect Claude Code

Add this MCP server to your Claude configuration:

{
  "mcpServers": {
    "agent-brain": {
      "type": "http",
      "url": "http://localhost:3100/mcp"
    }
  }
}

Restart Claude Code and run /mcp. You should see Agent Brain listed with its tools.

4. Store your first memory

Ask your agent to remember something durable:

Remember this: this project uses PostgreSQL for persistent storage and Redis only for cache. Tag it as an architecture decision.

Agent Brain stores the memory, embeds it, and makes it available to other connected agents.

5. Recall it from another session

Start a fresh session or switch to another connected agent:

What database do we use for persistent storage?

The agent can retrieve the memory without you re-explaining the decision.