GAIA

GAIA is AMD's open-source framework for building intelligent AI agents that run entirely on-device using AMD Ryzen AI hardware. It eliminates cloud costs and keeps data private, making it suitable for air-gapped, HIPAA-compliant, and GDPR-friendly deployments. The framework supports Python and C++17, and includes a full agent UI, RAG system, voice integration, and MCP connectivity. Install via pip install amd-gaia and build your first agent in minutes using the quickstart guide.

• Agent Framework — Base class with tool orchestration, state management, and error recovery for building custom agents.
• 100% Local Execution — All inference runs on-device using AMD Ryzen AI NPU + iGPU; no API fees or usage limits.
• Privacy-First Design — Data never leaves your machine; suitable for healthcare, finance, and enterprise air-gapped environments.
• RAG System — Document indexing and semantic search supporting Q&A over 50+ file formats.
• Voice Integration — Whisper ASR and Kokoro TTS enable full speech interaction with local models.
• Vision Models — Extract text from images using Qwen3-VL-4B for multimodal workflows.
• MCP Integration — Connect agents to any MCP server for external tool access and extensibility.
• Plugin System — Distribute and discover agents via PyPI with auto-discovery support.
• C++ Framework — A C++17 port of the base agent framework for native applications without Python dependencies.
• Agent UI — Privacy-first desktop app with chat, file browser, document indexing, and tool execution.