Original Discord Post by sacco_89595 | 2024-09-28 22:38:28
ChatGPT-4o’s voice bot can distinguish various voice tones, handle interruptions, and respond to inquiries in real-time. It’s trained to deliver more natural speech, adjusting its voice based on emotions. Currently, the Japanese voice options available on Convai sound somewhat unnatural. Is there a way to use the GPT-4o voice bot? If users can set it up themselves, please provide instructions.
Reply by madnomad4540 | 2024-09-29 13:21:51
Hello i have the same question, but if it can work, could it be used with elevenlabs voices?
Reply by sacco_89595 | 2024-10-03 00:37:53
<@&1163218672580575372> <@1023671043287699568> Hello, please answer me.
Reply by d_acharyya | 2024-10-03 05:05:31
Hi <@1170263861010628654> <@1086715236230377672> I believe you guys are referring to the OpenAI’s voice-to-voice realtime api. We are currently looking into it. We will soon provide an update on this.
Reply by sacco_89595 | 2024-10-03 05:09:21
That’s wonderful! Thank you very much. How soon is it ? Will it be implemented within this year, or next week?
Reply by d_acharyya | 2024-10-03 05:17:56
we are still in scoping out phase. Once this is done, we will be able to provide some timelines.
Reply by sacco_89595 | 2024-10-03 06:11:36
I’m really looking forward to it!
Reply by madnomad4540 | 2024-10-03 10:13:10
Thank you deepankar, it would be awesome
Reply by boomslanghegemony | 2024-10-03 13:05:18
That would be scary levels of realism. I honestly get the occasional chill talking to my characters. I can’t imagine how hard it must be to keep up with all the advances.
This conversation happened on the Convai Discord Server, so this post will be closed.