Creating text to speech audio with openai turbo4o API

Has anybody worked on creating audio with openai turbo4o API , for text-to-speech conversion with emotions?

You’re probably referring to the new “Advanced Voice Mode” that OpenAI demoed in May, which is capable of sounding very emotive. That voice mode is apparently going to start rolling out to a select group of users in late July and be available to all ChatGPT Plus customers in Fall.

Right now the only way would be to generate text with GPT-4o (or GPT-4-Turbo - keep in mind that these are two different models with different pricing and capabilities) to generate text, then pass that text to a TTS program like OpenAI’s or ElevenLabs. There’s no current way to control the emotion reliably.

Hi @praveenmenon999,

In case you’re referring to the gpt-4o - currently only text-to-text and vision capabilities are GA on the API for this model.

As of now, you can use the audio endpoint to create speech.