I’m trying to improve my AI agent’s response speed and make it sound more human. My current latency is around 840ms total: Transcriber (Deepgram): ~100ms Model: GPT-4o Mini ~390ms Voice (Cartesia): ~250ms Transport (Web): ~100ms Even with this, the agent still feels slow when responding. Any tips on reducing perceived latency or improving turn-taking / naturalness? Are there better configs, models, or streaming strategies you’d recommend? Thanks!

Hi RoH, yes we have a guide specifically on reducing latency here: https://vapi.ai/blog/optimize-voice-ai-performance-tips-and-strategies-with-vapi Thanks!

Hello Shaunak, thank you so much! I’ll go through the blog.

No worries! I'll mark this ticket as closed. Feel free to reach out to us if you have any further queries. Thanks, Shaunak