Aug 7
Support custom voice models on Mac and iOS (Mistral Voxtral)
Referring the Discord post.
There's now an option to connect Whisper voice models through Groq on macOS. I suppose this will also be available in the next version of iOS too.
At the same time, the French company Mistral released its own ASR models called Voxtral.
They are available through the API https://docs.mistral.ai/capabilities/audio/.
It would be great if these models could also be supported on Mac and iOS SW versions, especially for users who don't have powerful devices.
Mistral has lower latency for EU users compared to the US ASR cloud solutions.
Pending
I would be nice to have voice custom model integration (audio/transcriptions openai API compatible) similar to the current custom LLM Integration given several providers use this standard API