We are hence planning to develop this in-house. However if there is someone out there who provides low latency text-to-speech in conversational Hindi, then we would much rather use their solution.
Do you guys know anyone who does? Or do you think we should go after this ourselves?
https://news.microsoft.com/en-in/indian-startup-sarvam-ai-co...
https://fortune.com/asia/2024/01/27/microsoft-ai-india-langu...
https://news.microsoft.com/source/asia/features/microsoft-re...
I’d also check out https://github.com/dubverse-ai/MahaTTS
What are some examples of how these things differ? I've been exploring Hindi recently, but find that I'm learning some pretty stuffy speech from Snell's books.