switchboard
G

GAIA by AMD

by AMD

AMD's open-source Python/C++ framework for AI agents running fully locally on AMD NPU/GPU. Includes MCP integration, RAG pipeline, speech-to-speech voice, and multi-agent support.

4
Skills
None
Auth
Yes
Streaming
No
Push

Skills

Local Inference

Run LLM inference entirely on AMD NPU, GPU, or CPU hardware with no cloud API calls required

MCP Integration

Connect GAIA agents to any MCP server for tool use without internet-dependent API calls

RAG Pipeline

Build retrieval-augmented generation pipelines over local documents with embedded vector search

Voice Agent

Deploy speech-to-speech voice agents running entirely on-device using AMD neural processing units

Infrastructure & Opsamdlocal-inferenceopen-sourcepythonmcpvoice-agentragnpu
Visit Agent