OuteTTS

Why we included this project

OuteTTS is a small Python library that wraps the OuteTTS 1.0 models for on-device text-to-speech. A few lines of code are enough to generate audio and reuse saved speaker profiles, or clone a new voice from a short recording. The same interface runs on llama.cpp, Hugging Face Transformers, ExLlamaV2, and vLLM, with batched generation available for higher GPU throughput and sensible default sampling settings. A Transformers.js binding brings the same capability to browsers. It's a solid starting point for local speech synthesis, from quick prototypes to on-prem assistants, and it handles multilingual output from a compact model set.