openai_tts
Text-to-speech for Home Assistant from OpenAI, Mistral, Groq, Lemonfox, Kokoro, Chatterbox or any server that implements the OpenAI speech API, with announcemen
What is openai_tts?
Text-to-speech for Home Assistant from OpenAI, Mistral, Groq, Lemonfox, Kokoro, Chatterbox or any server that implements the OpenAI speech API, with announcements that restore your volume and music.
Overview of openai_tts
What it is
OpenAI TTS is a text-to-speech integration for Home Assistant that works with any service implementing the OpenAI speech API. It supports cloud providers such as OpenAI, Mistral, Groq, Lemonfox, and OpenRouter, as well as self-hosted servers like Kokoro-FastAPI, Chatterbox, or any custom endpoint. The integration is designed to handle announcements that restore the previous volume and music playback after the message finishes.
Key features
- Works with any server that implements the OpenAI speech API, including cloud and self-hosted options.
- Presets for OpenAI, Mistral, Groq, Lemonfox, Kokoro, Chatterbox, OpenRouter, and Custom endpoints fill in the endpoint, models, and voices.
- Voice lists are read live from providers that publish them, or from a fixed list for those that do not.
- Announcements restore the volume and music after playing, with support for volume override on Sonos and Music Assistant.
- Gain control per profile from -12 to +12 dB in 0.5 dB steps, with a limiter to prevent clipping.
- API keys are validated when an entry is created, re-entered, or via the
openai_tts.set_api_keyaction.
Use cases
- Home automation announcements that pause and resume music or other audio.
- Voice assistant responses using the Home Assistant Assist pipeline.
- Multi-room audio announcements with Music Assistant groups, where the integration keeps the group playing.
- Custom text-to-speech for any Home Assistant automation or script.
Getting started
See the repository README for installation and configuration instructions.
License and pricing
The license is GPL-3.0. The integration is open source and has no stated pricing.
Based on the project's public README and repository metadata.