Local Setup

CUDA Setup (For NVIDIA GPUs)

CUDA speeds up local inference significantly. Ensure you’ve installed the CUDA toolkit, including cuDNN and cuBLAS.

npm install npx --no

node-llama-cpp download --gpu cuda

In addition to the basic setup, configure OpenAI and ElevenLabs API keys for extended cloud functionalities.

Adjust these for voice stability and style: ELEVENLABS_VOICE_STABILITY=0.5

ELEVENLABS_VOICE_SIMILARITY_BOOST=0.9

ELEVENLABS_VOICE_STYLE=0.66

Last updated 8 months ago