Orchestration Models
All the fancy stuff Vapi does on top of the core models.
Vapi also runs a suite of audio and text models that make it’s latency-optimized Speech-to-Text (STT), Large Language Model (LLM), & Text-to-Speech (TTS) pipeline feel human.
Here’s a high-level overview of the Vapi architecture:
![](https://mintlify.s3-us-west-1.amazonaws.com/vapi/static/images/learn/platform/vapi-orchestration.png)
Support and Optimization
To provide you and your customers with a superior conversational experience, we have various latency optimizations like end-to-end streaming and colocating servers that shave off every possible millisecond of latency. We also manage the coordination of interruptions, turn-taking, and other conversational dynamics.
Customizable
We built-in many smaller features to give developers a lot of room to customize and integrate. For example, there’s no need for you to hook up Twilio websockets or build bidirectional audio streaming. Instead, you can connect to the WebRTC stream through our Web, iOS, or Python clients…and then get right back to what you were doing.
Voice AI at Any Scale
Finally, we designed Vapi to be highly scalable. We accommodate everything from small businesses and companies all the way up to enterprise-level clients.
Was this page helpful?