Vapi is the Voice AI platform for developers.

Vapi AI

Call id 37676589-2a83-4cf1-88ab-cb6368247c8f

01:42:40:725
[LOG]
Voice input: Could you please describe the situation with the poking wire in more detail?

01:42:40:725
[LOG]
Model output: Could you please describe the situation with the poking wire in more detail? <flush/>

01:42:41:183
[LOG]
Model request cost (attempt #1, $0, 245 prompt, 2 completion)

You can continue this conversation by mentioning me in the message.

<@1233396054796275854>
Please provide me with the context you would like to analyze for the call with ID 37676589-2a83-4cf1-88ab-cb6368247c8f. Would you like a cost breakdown, the transcript, or the recording URLs?

the 11labs ws-connection was closed for** "Yes. I can hear you. Is the poking wire causing an emergency situation for you?"** before it was voiced out because of user input arrived and another llm request was sent again and then later on user input **"Hello?**" it was voiced out.

<@865531204849500190> could we do something to mandatory pronounce some chunks, no matter if user interrupt

as of now there is no feature for `force-say` few of llm output.