Skip to main content
Run GAIA on your own machine with the GAIA CLI. It scaffolds your environment, starts every service with Docker, and gets you to a working instance in a few commands.

​
What runs

  • Web: Next.js
  • Backend: FastAPI
  • Databases/Infra: PostgreSQL, MongoDB, Redis, ChromaDB, RabbitMQ (via Docker)
  • Workspace storage: JuiceFS (per-user file uploads, agent artifacts, skills)

​
Prerequisites

  • Node.js 20+ and one package manager (npm, pnpm, or bun)
  • Docker Engine 20.10+ and Compose v2+
  • 2+ CPU cores, 4GB+ RAM, 10GB+ disk (8GB+ RAM recommended)
  • An API key for at least one supported LLM provider (OpenAI or Google Gemini), see Environment variables

​
Quick start

1

Install the GAIA CLI

2

Initialize self-host mode

gaia init
In the setup wizard, choose Self-Host (Docker).
3

Start services

gaia start
4

Monitor and open GAIA

gaia status
gaia logs

​
Advanced: workspace storage on macOS and Windows

You only need this if you plan to run the backend outside the default Docker setup.
Run the backend in a container on every platform. Its workspace storage (file uploads, artifacts, skills) uses JuiceFS, which mounts through Linux FUSE, a native backend on macOS or Windows can’t provide it. Docker Desktop covers this on all platforms. See Docker setup for the FUSE requirements and configuration.

​
Next steps