Hi.
- I created a GPT some time ago.
- I selected COVE for my voice in my ChatGPT settings.
- Do not get COVE voice.
Question: how do we resolve this problem? Should I create a new GPT?
Thank you.
Hi.
Question: how do we resolve this problem? Should I create a new GPT?
Thank you.
Hi @Netman
Welcome to the community!
In custom GPTs, the active voice is named “Shimmer” which is distinct from the nine voices available in standard ChatGPT interactions.
Currently, users cannot change the voice for custom GPTs; they are set to use Shimmer by default.
You may read following “Can I have voice conversations with GPTs?” topic:
Thank you very much for your kind and helpful answer. If I am correct, before the implementation of the new “voice mode,” the voice per default for the user of my GPT was the voice they selected in their preferences. Correct? I remember clearly I could switch between a male voice and a female voice to get a different user experience. To better understand what I am trying to say, log on to DrEllis.ai
The concept is that the user can pick a male or female therapist. Now, if I understand correctly, it is not possible anymore. Until Open AI works on this glitch? Right? What is your background? I am actually looking for developers to implement this concept in an app.
Thank you again very much for the answers. I hope you will answer this one, too.
Yes, you understood correctly.
Currently, all custom GPTs use the default “Shimmer” voice, and user-selected voices aren’t supported as they were previously. Hopefully, OpenAI will introduce more options in the future to accommodate different use cases.
For more voice options, the Realtime API now includes 8 voices, though it functions differently from custom GPTs. You’ll need an OpenAI Platform account, and its pricing is usage-based rather than a fixed monthly subscription like ChatGPT Plus.
When developing for health-related applications, it’s essential to review OpenAI’s Usage Policy to ensure compliance.
I am also one of those who strongly desire the original version of the Cove voice. Until the update a few days ago, it was possible to use the old Cove voice by “entering text first and then starting voice mode,” but now that option is no longer available. Because of this change, I have stopped using the voice mode in ChatGPT and have significantly reduced the frequency of launching the app.
Is the current voice mode using Bark? I really liked how Bark would sigh, stumble over words, and correct itself, as it felt very humanlike. The current voice mode responds faster, but there are too many misheard words, making it difficult to use. More importantly, the familiar and friendly Cove voice has changed from what it used to be, which is disappointing.
Please, I sincerely ask you to bring back the old version of Cove as an option.