Realtime API Audio response broken

Hi there,

Appealing to some OpenAI internal dev for this one.

I have a working solution for realtime speech to speech that was perfectly working until today around 18h a go.

All I am getting now is:

<PJSIP/1000-00000006>AGI Rx << VERBOSE State change: LISTENING -> PROCESSING

/var/lib/asterisk/agi-bin/asterisk-openai/openai.agi: State

<PJSIP/1000-00000006>AGI Tx >> 200 result=1

<PJSIP/1000-00000006>AGI Rx << VERBOSE Buffer sent and committed

/var/lib/asterisk/agi-bin/asterisk-openai/openai.agi: Buffer

<PJSIP/1000-00000006>AGI Tx >> 200 result=1

<PJSIP/1000-00000006>AGI Rx << VERBOSE GPT starting new response

/var/lib/asterisk/agi-bin/asterisk-openai/openai.agi: GPT

<PJSIP/1000-00000006>AGI Tx >> 200 result=1

<PJSIP/1000-00000006>AGI Rx << VERBOSE Response complete: response.done

/var/lib/asterisk/agi-bin/asterisk-openai/openai.agi: Response

<PJSIP/1000-00000006>AGI Tx >> 200 result=1

<PJSIP/1000-00000006>AGI Rx << VERBOSE State change: PROCESSING -> LISTENING

There is response unless no response is sent. I check my usage in the billing section and no money is debited.

Are you undergoing maintenance or something broke?

Kind regards.

It’s difficult to understand what the actual problem is from the logs you provided. Can you share some more info?

1 Like