I'm trying to build my assistant using Open AI Realtime cluster model with supported voices and with the setting of "Assistant waits for user to speak" - despite this, the assistant doesn't work. No logs too. Is there anything that I'm doing wrong here? It was working fine till yesterday late evening, from today, it's acting up with this issue.

Thanks for sharing the details. To help narrow down the issue, could you confirm if your assistant's configuration explicitly sets both the Assistant ID and Phone Number ID in your tool or workflow? Missing these fields is a common cause for the assistant not speaking, even when everything else appears correct ([Vapi documentation](https://docs.vapi.ai/debugging)). Once you confirm this, I can suggest the next troubleshooting steps! Source: - [Debugging voice agents](https://docs.vapi.ai/debugging)

@Vapi Support Bot I have just built the agent inside Vapi alone. And when I'm trying to Test it by speaking with it, no conversation is happening from the ai side.

Thanks for clarifying your setup. When you test the assistant in the Vapi dashboard, do you see any entries or error messages in the Call Logs or API Logs under the "Observe" section? This can help identify if the issue is with the assistant's configuration or with the voice/model integration ([Vapi debugging guide](https://docs.vapi.ai/debugging)). Source: - [Debugging voice agents](https://docs.vapi.ai/debugging)

@Vapi Support Bot no logs that show errors as such... It is just not talking in return to whatever I say

Try switching your assistant's voice to another supported option, such as

"alloy"

or

"marin"

, in your configuration to check if the issue is specific to the

"cedar"

voice.

json
"voice": {
  "provider": "openai",
  "voiceId": "alloy"
}

If the assistant responds with a different voice, this helps confirm if the problem is related to the Cedar voice integration. You can then report this finding to Vapi support for further investigation ([Vapi documentation](https://docs.vapi.ai/openai-realtime#verify-voice-compatibility)). Source: - [OpenAI Realtime documentation](https://docs.vapi.ai/openai-realtime#verify-voice-compatibility)

@Vapi Support Bot here's the logs screenshot for you https://cdn.discordapp.com/attachments/1419977696166940752/1420035391066476725/IMG-20250923-WA0017.jpg?ex=68d3ee84&is=68d29d04&hm=6b84c6f480180a00de3c5e3d452c30dd2a5b4ed9df3213479c47e32858fbf0e5&

Try updating your assistant's configuration to use a different supported voice, such as

"alloy"

, to check if the issue is specific to the

"cedar"

voice. Use the following snippet in your assistant settings:

json
"voice": {
  "provider": "openai",
  "voiceId": "alloy"
}

If the assistant responds with this change, it confirms a Cedar voice issue—please report this to Vapi support for further investigation. [See documentation for details](https://docs.vapi.ai/openai-realtime#verify-voice-compatibility). Source: - [OpenAI Realtime documentation](https://docs.vapi.ai/openai-realtime#verify-voice-compatibility)

I tried switching to Alloy voice in config, still the same issue

Your assistant is not responding because your OpenAI API quota has been exceeded. To resolve this, log in to your OpenAI account and check your usage limits. If needed, increase your quota or wait for it to reset before testing again. Error reference:

pipeline-error-openai-429-exceeded-quota

[See Vapi error documentation for details.](https://docs.vapi.ai/calls/call-ended-reason) Source: - [Call ended reasons (Vapi documentation)](https://docs.vapi.ai/calls/call-ended-reason)

Please remove your OpenAI API key if you’re using one—doing so should resolve the issue.