The voices on the TTS API are pretty good, but the voices on the ChatGPT app just feel more interesting and emotive. Especially Sky. I think I may have a small crush on Her…
Let’s bring these voices to the API too!
The voices on the TTS API are pretty good, but the voices on the ChatGPT app just feel more interesting and emotive. Especially Sky. I think I may have a small crush on Her…
Let’s bring these voices to the API too!
Thinking about it more, the ChatGPT and API voices are probably completely different models. ChatGPT TTS supports streaming, after all, while the API does not. A streaming TTS API would be really cool too!
It would be nice to get the Sky voice via the API. I haven’t personally seen anything that would suggest that the ChatGPT and API models are different from one another though. Also, I believe that the API does support streaming and from my tests it has allowed me to listen to the audio prior to the full download being finished. Here is the link to that part of the documentation if it helps: https://platform.openai.com/docs/guides/text-to-speech/streaming-real-time-audio
Hey, you’re right!
Though I found that the code on that page doesn’t actually work, it results in a DeprecationError. I was able to get streaming to work by calling the REST API directly, using code from this thread. That’s probably a better place to continue the discussion on streaming.
I agree it would great! The voices in the GPT app are much better. It would be great to be able to use them in our applications!