Hello Team, This is my configuration: - custom-transcriber - custom-LLM - custom-voice VAPI send me many silence time out. I don´t kown whats is the reason? Can you help me please? This is the callId: - 019a0ace-7a87-7886-80d5-9d2c24d5dd97

To diagnose the silence timeouts with your custom setup, go to your [Vapi dashboard](https://dashboard.vapi.ai/), then navigate to

Observe > Call Logs

and search for your call ID (

019a0ace-7a87-7886-80d5-9d2c24d5dd97

). Review the call transcript and tool execution details to pinpoint where the delay or silence occurred. This will help you identify if the issue is with your custom transcriber, LLM, or voice integration. Tip: Focus on any long processing times or missing responses in the log to guide your debugging. [See Debugging Guide](https://docs.vapi.ai/debugging) Source: - [Debugging voice agents](https://docs.vapi.ai/debugging)

Another call: - 019a16a4-5830-7bb3-809f-36549385ab28

Hi aivini, To configure your assistant to say something before ending a call due to silence, you can use idle messages triggered by the

customer.speech.timeout

hook. Here is a simplified configuration example:

json
{
  "hooks": [
    {
      "on": "customer.speech.timeout",
      "options": { "timeoutSeconds": 30, "triggerMaxCount": 1, "triggerResetMode": "onUserSpeech" },
      "do": [
        { "type": "say", "exact": "I'll be ending the call now, please feel free to call back at any time." },
        { "type": "tool", "tool": { "type": "endCall" } }
      ]
    }
  ]
}

- Set

timeoutSeconds

to the desired period of silence. - Configure

triggerMaxCount

to limit how often the message is triggered. - Use

triggerResetMode

to specify if the trigger count should reset when the user speaks. You can find detailed configuration guidance in the [idle messages documentation](https://docs.vapi.ai/assistants/idle-messages).

But de problem is: ¿Why VAPI don't send me the chat/completions? callId: 019a1ac1-6e84-7bb8-9a99-13aec5a7a93d This response from the cliente:

Alcala

never VAPI send me. ¿Is turn taking broke? thanks https://cdn.discordapp.com/attachments/1430837260722376734/1431682918765564094/image.png?ex=68fe4e1f&is=68fcfc9f&hm=6d4e60fd74f0791b667ff6f134b8e723260cc3112bf33bdd82c94c3995700929&

Hi aivini, We checked the logs and it looks like your custom-voice implementation is failing on our platform. Here are the relevant entries:

🟡 09:46:24:251 Couldn't Get Speech For `custom-voice:https://api-channel-voice-188691320542.europe-southwest1.run.app/api/synthesize:16000:0e417dbcdda08112b453431be24286b1c85a443e566d62c1c95469d355b1d7b0`, "Perdona, no te entendí. ¿Podrías repetirlo?"

🟡 09:46:24:253 Couldn't Complete Say Request {
  "task": {
    "content": "Perdona, no te entendí. ¿Podrías repetirlo?"
  },
  "error": {}
}

What’s happening is: - The assistant sends a

say

request. - The speech cannot be generated by your custom-voice service. - As a result, nothing is played back to the caller and the call eventually times out due to silence. We recommend switching to a stock voice from one of the supported providers (e.g., ElevenLabs, Azure, or another provider listed in our [API documentation](https://docs.vapi.ai/api-reference/voices)) to avoid this issue. Let us know if you’d like help configuring a stock voice, or if you plan to debug your custom-voice service further. Best, VAPI Support ---