I switched to 4.1  from 4o and it's worst. All my customers are complaining. I'm considering churning from Vapi but would love to figure this out.  Can you help me figure out how to adjust my flow to make it better for latency & voice quality? My setup is - LLM: 4.1 - STT: Deepgram nova-3 - TTS: ElevenLabs Multilingual V2 - We tested smartEndpointingPlan with the LiveKit provider (waitFunction: "700 + 4000 * max(0, x-0.5)", waitSeconds: 0.4) but it's still slow. Thank you so much, Yan

To reduce latency, try making your smart endpointing plan more aggressive. For example, update your configuration as follows:

json
"smartEndpointingPlan": {
  "provider": "livekit",
  "waitFunction": "200 + 8000 * x",
  "waitSeconds": 0.2
}

This should help your assistant respond faster—test and adjust the values as needed for your use case. For more details, see the voice pipeline configuration documentation. Source: - [Voice pipeline configuration](https://docs.vapi.ai/customization/voice-pipeline-configuration)

even without that smart endpointing its slow

can you please share your call id so we can investigate