isaiahbjork/orpheus-tts-local

Provides a Python client for the Orpheus 3B TTS model with eight distinct voices and emotion tags (giggle, laugh, sigh, etc.) for expressive speech synthesis. Communicates with LM Studio's local inference server via its API, supporting configurable generation parameters like temperature, top-p sampling, and repetition penalty. Outputs directly to WAV files with command-line control over voice selection and generation behavior.

525 stars. No commits in the last 6 months.