Create a specific voice for the app

I would like one of my GPT models to have a specific voice for the app.

Is it possible to choose more than the five available voice patterns in the app, or to prompt the GPT to use a specific voice when interacting with users through the app?

No. The voice patterns available in the app are the only voices available.

You can choose one of the voices GPT has available when you build a custom GPT (like saying “hey, could you make the voice for this GPT into “cove” please?”), but that is the only thing currently available right now.

If you want a custom voice generated, you would have to build an Assistant using the API, as well as an API from another resource that would have the voice you want to use (assuming you don’t like OpenAI’s current TTS options).

This is still a new feature that’s in it’s infancy. I’m sure customization will continue and evolve over time, and relatively soon too. However, they might be extra careful with this so long as to not accidentally create a deepfake generator.

Thanks for the detailed answer Macha ! It makes sense.
I was asking this question because Thomas Dimson showed how he built a GPT with all the features of a “salty pirate” and he succeed to change the voice for the app to something in adequation of his GPT.

So, or he used an API and I didn’t pay attention or it was a demo for the next gen of GPT …
Here it is from 3m56 till 9m

1 Like