How to control the text to speech?

The documentation says that for the gpt-4o-mini-tts we can control different aspects

using the instructions prop of the API.

  • Accent
  • Emotional range
  • Intonation
  • Impressions
  • Speed of speech
  • Tone
  • Whispering

However, there are no examples at all… is it a simple prompt?
Does anyone already toy with this?

There are many examples in this live demo:

There is a toggle in the upper right that gives you the equivalent code for each example.

You can also find all the example prompts here:

2 Likes

OK Thank you I’l have a look :folded_hands:

OK So I had a look.
and indeed it’s much clearer.

Are the instructions like Affect, Tone etc… free?
I mean, it’s just prompting, I guess?

Free… style. But if you meant costs, they are part of text input tokens pricing.