So we have created an outbound sales assistant that is going to call a list of leads. Testing this by calling myself and the delay that we are having from the initial pick up from what would be the customer and them saying hello to the agent replying is 2-4 seconds. This is unusable and genuinely not something that we could even deploy. Is there some secret setting that we need to make the reponses less than 1s 🤣 https://cdn.discordapp.com/attachments/1463126187390926952/1463126187541790773/image.png?ex=6979ec72&is=69789af2&hm=3b612990703d235978632cbb050383aa60b4f1040cdfefd6aae1d81b5d3f08b4&

Thank you for sharing the details. To help pinpoint the cause of the 2–4 second delay, could you let me know which speech-to-text, language model, and text-to-speech providers you are currently using in your voice pipeline (e.g., OpenAI Realtime, Deepgram, ElevenLabs, etc.)? This will help determine if more aggressive settings or a different provider could reduce your response time closer to 1 second. If you have a sample of your current pipeline configuration (such as your

startSpeakingPlan

or

waitFunction

), sharing that would also be helpful for more specific guidance. Sources: - [Vapi Voice Pipeline Configuration Documentation](https://docs.vapi.ai/customization/voice-pipeline-configuration) - [Vapi Speech Configuration Documentation](https://docs.vapi.ai/customization/speech-configuration) - [Vapi OpenAI Realtime Documentation](https://docs.vapi.ai/openai-realtime)

"startSpeakingPlan": { "waitSeconds": 0.1 "model": { "model": "gpt-4o-mini", "voiceId": "Mia", "provider": "vapi"

Please share call Id that's having this issue.

019bdae6-3bf9-7aa9-b056-8c4544d44966 019bdaef-3608-7116-aefa-c79b83a943f3

Hey Dale, the problem is originating from the provider side; there's no such latency found in logs. It means we are sending data at the correct time, but telnyx taking extra time to deliver to you. Can you try explore Twilio as provider for now?