Orpheus-FastAPI

Why we included this project

Orpheus-FASTAPI is a self-hosted text-to-speech server that exposes an OpenAI-compatible API, so if you already talk to /v1/audio/speech, it slots in without rewriting your client. It wraps the Orpheus model in a FastAPI service and includes a web UI with waveform playback for generating audio without writing code. The multilingual support is the most useful part: 24 voices across eight languages, with language-specific quantized GGUF models you swap by editing one environment variable. Emotion tags let you nudge delivery toward laughter, sighs, or other expressive cues, and long-form text gets sentence batching with crossfade stitching to keep output seamless. Docker Compose covers CUDA, ROCm, and CPU-only setups, so it works on an RTX card or a modest box.