Apocrypha

amd-gaia

Lightweight agent framework for the edge and AMD Ryzen AI PCs

GAIA is AMD's open-source agent framework for local AI agents on Ryzen AI hardware (NPU + iGPU). It orchestrates LLM-driven workflows over any OpenAI-compatible inference endpoint, with built-in integrations for Docker, Jira, code-search, RAG, MCP servers, and Whisper / Kokoro voice pipelines. The reference local backend is Lemonade Server (sci-ml/lemonade); GAIA itself is hardware-agnostic so long as the upstream LLM API is OpenAI-compatible.

Available in

OverlayNewestEbuildsLast activity
stuff GitHub ↗ 0.22.0 2 2 d details ›

Versions & arches

VersionOverlay amd64 Committed
0.22.0 stuff amd64 testing 10 d view · download · history ↗
0.21.2-r1 stuff amd64 testing 10 d view · download · history ↗

Use flags of 0.22.0

  • +api Install FastAPI / uvicorn HTTP-server bindings (gaia.api)
  • audio Pull sci-ml/pytorch for the gaia.audio module (Whisper ASR / Kokoro TTS prereq; upstream's audio extra also lists torchvision/torchaudio but gaia code never imports them)
  • +mcp Install Model Context Protocol client/server support (gaia.mcp)
  • eval Install evaluation harness deps (anthropic, scikit-learn, reportlab) for gaia.eval benchmarks
  • image Pull dev-python/term-image for terminal-rendered image output (gaia.image)
  • talk Pull openai-whisper (ASR) + sounddevice (audio I/O) for gaia.talk's voice pipeline; the Kokoro TTS half is unavailable until that chain lands and is not pulled by this flag
  • ui Build the gaia.ui web frontend (FastAPI + faiss-backed RAG over PDFs via PyMuPDF + sentence-transformers); requires +api