Docker containers for AMD GAIA - the AI framework for building and deploying intelligent agents.
Current Versions: gaia-linux: 1.0.0 | gaia-dev: 1.1.0
This project provides production-ready Docker containers for AMD GAIA in multiple variants, each optimized for different use cases.
| Container | Use Case | Installation Method | Documentation |
|---|---|---|---|
| gaia-linux | Production runtime | PyPI at startup | View Docs |
| gaia-dev | Development + Claude Code | Cloned from source | View Docs |
| gaia-windows | Windows environments | Planned | View Docs |
Choose the container that matches your needs:
Use gaia-linux for running GAIA applications:
docker pull itomek/gaia-linux:1.0.0
# Installs latest GAIA from PyPI automatically
docker run -dit \
--name gaia-linux \
-e LEMONADE_BASE_URL=https://your-server.com/api/v1 \
itomek/gaia-linux:1.0.0
# Or pin a specific GAIA version
docker run -dit \
--name gaia-linux \
-e LEMONADE_BASE_URL=https://your-server.com/api/v1 \
-e GAIA_VERSION=0.15.3.2 \
itomek/gaia-linux:1.0.0See gaia-linux documentation for complete usage.
Use gaia-dev for developing GAIA or contributing to the project:
docker pull itomek/gaia-dev:1.1.0
docker run -dit \
--name gaia-dev \
-v gaia-src:/home/gaia/gaia \
-e LEMONADE_BASE_URL=https://your-server.com/api/v1 \
-e GAIA_REPO_URL=https://github.com/amd/gaia.git \
-e GITHUB_TOKEN=ghp_your_token \
itomek/gaia-dev:1.1.0See gaia-dev documentation for complete usage.
- Decoupled versioning: Image versions track the base environment; GAIA version is chosen at runtime
- Fast installation: Uses
uvpackage manager for 10-100x faster installs than pip - Complete isolation: Each container is fully self-contained
- Multi-instance support: Run multiple GAIA instances on the same host
- Production-ready: Built on Ubuntu 24.04 LTS with security best practices
- Best for: Running GAIA applications, production deployments
- GAIA source: Installed from PyPI at startup (latest or pinned version)
- Includes: Ubuntu 24.04, Python 3.12, Node.js 20, system dependencies
- Size: Smaller (~1-2 GB)
- Startup: 2-3 minutes first run, 30 seconds cached
- Best for: GAIA development, contributions, experimentation
- GAIA source: Cloned from GitHub, editable install
- Includes: Everything in gaia-linux + Claude Code, network isolation tools
- Size: Larger (~2-3 GB)
- Startup: 3-5 minutes first run, 1 minute cached
- Best for: Windows-specific workflows
- Status: Future feature, in planning phase
- Alternative: Use gaia-linux or gaia-dev with Docker Desktop on Windows
- gaia-linux Documentation - Runtime container guide
- gaia-dev Documentation - Development container guide
- gaia-windows Documentation - Windows container (planned)
- Using as Base Image - Extend containers in your own Dockerfile
- Building from Source - Build containers locally
Docker image versions are managed in VERSION.json. These versions track the base environment (Ubuntu + Python + Node.js + system deps), not the GAIA PyPI package version. GAIA is installed at runtime, so a new GAIA release does not require a new Docker image.
# Pull image (version tracks base environment)
docker pull itomek/gaia-linux:1.0.0
docker pull itomek/gaia-dev:1.1.0
# Use latest GAIA (default)
docker run -dit -e LEMONADE_BASE_URL=... itomek/gaia-linux:1.0.0
# Pin specific GAIA version
docker run -dit -e LEMONADE_BASE_URL=... -e GAIA_VERSION=0.15.3.2 itomek/gaia-linux:1.0.0When the base environment changes:
- Update
VERSION.json - CI automatically builds and publishes new images
- GitHub release created with version tag
# Install dependencies
uv sync
# Run all tests
uv run pytest tests/ -v
# Run specific test suite
uv run pytest tests/test_dockerfile.py -v# Build gaia-linux
docker build -f gaia-linux/Dockerfile -t itomek/gaia-linux:1.0.0 .
# Build gaia-dev
docker build -f gaia-dev/Dockerfile -t itomek/gaia-dev:1.1.0 .See CLAUDE.md for complete development documentation.
- GAIA Framework Issues: https://github.com/amd/gaia/issues
- Docker Container Issues: https://github.com/itomek/gaia-docker/issues
- GAIA Documentation: https://github.com/amd/gaia
Contributions welcome! Please:
- Read CLAUDE.md for project structure and development workflow
- Open an issue to discuss significant changes
- Submit pull requests with tests
MIT License - see LICENSE file
- AMD GAIA - The AI framework these containers package
- Claude Code - AI coding assistant included in gaia-dev
Docker Hub: