lhl
Published © Apache-2.0

voicechat2: Local AI Voice Chat

A fast AI voicechat system that runs entirely locally (SRT/LLM/TTS) via WebSockets. Compatible with AMD GPUs.

IntermediateFull instructions provided1 hour2,052

Things used in this project

Hardware components

AMD Radeon Pro W7900 GPU
AMD Radeon Pro W7900 GPU
×1

Software apps and online services

AMD ROCm™ Software
AMD ROCm™ Software
whisper.cpp
llama.cpp
Coqui TTS

Story

Read more

Code

voicechat2

See the README for install instructions

Credits

lhl

lhl

1 project • 7 followers
Hacking on web/scaling, geo, mobile, imaging, VR/XR, dataviz, hw/mcu (+custom pcbs), biomed, currently focused on Gen AI/LLMs

Comments