<description>&lt;p&gt;Nikhil Gupta is the cofounder and CTO of VAPI. He’s been in the trenches building and scaling one of the biggest voice platforms in the world.&lt;/p&gt;&lt;p&gt;On this episode, he explains how VAPI aims to create a voice-default future where we talk to all our computers—and goes deep into every step of VAPI’s voice pipelines and the technical challenges along the way.&lt;/p&gt;&lt;p&gt;Some highlights:&lt;/p&gt;&lt;p&gt;☎️ &lt;strong&gt;Massive scale: &lt;/strong&gt;VAPI has processed 44 million voice calls on their platform, growing from a COVID-era one-click Zoom meeting button to a full voice infrastructure company used by thousands of developers.&lt;/p&gt;&lt;p&gt;⚡️ &lt;strong&gt;Latency matters:&lt;/strong&gt; Voice AI needs to respond within 500 milliseconds to feel natural to humans. That means cleaning audio, detecting when users are done speaking, transcribing text, generating responses, converting to speech, and handling interruptions—all within a fraction of a second.&lt;/p&gt;&lt;p&gt;🗣️&lt;strong&gt;Voice-first future: &lt;/strong&gt;Nikhil is betting a future where voice becomes our default interface with all computing systems.&lt;/p&gt;&lt;p&gt;If you’ve ever wondered how voice API actually works—this is the episode for you. &lt;/p&gt;&lt;p&gt;&lt;strong&gt;Chapters&lt;/strong&gt;&lt;/p&gt;&lt;p&gt;00:00 - Introducing VAPI&lt;/p&gt;&lt;p&gt;04:42 - Pivoting through COVID&lt;/p&gt;&lt;p&gt;05:42 - ChatGPT existential crisis&lt;/p&gt;&lt;p&gt;08:33 - Technical challenges of voice&lt;/p&gt;&lt;p&gt;12:42 - Anatomy of a voice call&lt;/p&gt;&lt;p&gt;14:46 - Knowing when someone is done speaking&lt;/p&gt;&lt;p&gt;18:37 - Routing to the fastest model&lt;/p&gt;&lt;p&gt;22:07 - Knowledge and context injection&lt;/p&gt;&lt;p&gt;26:47 - The text-to-speech bottleneck&lt;/p&gt;&lt;p&gt;31:14 - Handling interruptions gracefully&lt;/p&gt;&lt;p&gt;33:43 - The 500-millisecond barrier&lt;/p&gt;&lt;p&gt;36:56 - The DNS latency discovery&lt;/p&gt;&lt;p&gt;39:25 - Scaling the team and what's next&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Links&lt;/strong&gt;&lt;/p&gt;&lt;p&gt;* &lt;a target="_blank" href="https://vapi.ai/"&gt;VAPI&lt;/a&gt; &lt;/p&gt;&lt;p&gt;* &lt;a target="_blank" href="https://www.linkedin.com/in/nikhilro/"&gt;Nikhil Gupta on LinkedIn&lt;/a&gt;&lt;/p&gt;&lt;p&gt;* &lt;a target="_blank" href="https://www.producthunt.com/products/vapi"&gt;VAPI on Product Hunt&lt;/a&gt;&lt;/p&gt;&lt;p&gt;* &lt;a target="_blank" href="https://stratechery.com/"&gt;Stratechery by Ben Thompson&lt;/a&gt; - Recommended reading from Nikhil&lt;/p&gt; &lt;br/&gt;&lt;br/&gt;This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit &lt;a href="https://www.smartspeakers.fm?utm_medium=podcast&amp;#38;utm_campaign=CTA_1"&gt;www.smartspeakers.fm&lt;/a&gt;</description>

Smart Speakers Podcast

Dave Zohrob & Harish Agarwal

The 500ms Dash—Nikhil Gupta, VAPI

APR 7, 202543 MIN
Smart Speakers Podcast

The 500ms Dash—Nikhil Gupta, VAPI

APR 7, 202543 MIN

Description

<p>Nikhil Gupta is the cofounder and CTO of VAPI. He’s been in the trenches building and scaling one of the biggest voice platforms in the world.</p><p>On this episode, he explains how VAPI aims to create a voice-default future where we talk to all our computers—and goes deep into every step of VAPI’s voice pipelines and the technical challenges along the way.</p><p>Some highlights:</p><p>☎️ <strong>Massive scale: </strong>VAPI has processed 44 million voice calls on their platform, growing from a COVID-era one-click Zoom meeting button to a full voice infrastructure company used by thousands of developers.</p><p>⚡️ <strong>Latency matters:</strong> Voice AI needs to respond within 500 milliseconds to feel natural to humans. That means cleaning audio, detecting when users are done speaking, transcribing text, generating responses, converting to speech, and handling interruptions—all within a fraction of a second.</p><p>🗣️<strong>Voice-first future: </strong>Nikhil is betting a future where voice becomes our default interface with all computing systems.</p><p>If you’ve ever wondered how voice API actually works—this is the episode for you. </p><p><strong>Chapters</strong></p><p>00:00 - Introducing VAPI</p><p>04:42 - Pivoting through COVID</p><p>05:42 - ChatGPT existential crisis</p><p>08:33 - Technical challenges of voice</p><p>12:42 - Anatomy of a voice call</p><p>14:46 - Knowing when someone is done speaking</p><p>18:37 - Routing to the fastest model</p><p>22:07 - Knowledge and context injection</p><p>26:47 - The text-to-speech bottleneck</p><p>31:14 - Handling interruptions gracefully</p><p>33:43 - The 500-millisecond barrier</p><p>36:56 - The DNS latency discovery</p><p>39:25 - Scaling the team and what's next</p><p><strong>Links</strong></p><p>* <a target="_blank" href="https://vapi.ai/">VAPI</a> </p><p>* <a target="_blank" href="https://www.linkedin.com/in/nikhilro/">Nikhil Gupta on LinkedIn</a></p><p>* <a target="_blank" href="https://www.producthunt.com/products/vapi">VAPI on Product Hunt</a></p><p>* <a target="_blank" href="https://stratechery.com/">Stratechery by Ben Thompson</a> - Recommended reading from Nikhil</p> <br/><br/>This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit <a href="https://www.smartspeakers.fm?utm_medium=podcast&#38;utm_campaign=CTA_1">www.smartspeakers.fm</a>