AhaTTS - Kokoro OpenAI-API

Language: English | 简体中文

AhaTTS is an open-source, production-ready text-to-speech service with an OpenAI-compatible API and a built-in web demo.

Highlights

• OpenAI-compatible API for drop-in integration
• Built-in api/web demo UI
• CPU / CUDA / MPS support
• Multi-language and multi-voice (Kokoro)
• Streaming output with high-quality audio

Quickstart

./scripts/install.sh --device cpu   # or gpu / mac
./scripts/dev.sh

Visit:

• http://localhost:25288/docs
• http://localhost:25288/web/

OpenAI Speech API

Request URL: http(s)://<server-address>:<port>/v1/audio/speech (POST).

Parameters:

• model (string, required): one of tts-1 or tts-1-hd.
• input (string, required): text to generate audio from. Max length 4096 characters.
• voice (string, required): one of alloy, ash, coral, echo, fable, onyx, nova, sage, shimmer.
• response_format (string, optional): audio format, default mp3. Supported: mp3, opus, aac, flac.
• speed (number, optional): audio speed, default 1.0. Range 0.5 to 2.0.

Acknowledgements

Special thanks to the following projects for inspiration and reference implementations:

• Kokoro-FastAPI: https://github.com/remsky/Kokoro-FastAPI
• kokoro: https://github.com/hexgrad/kokoro