Vapi Made Crystal Clear · chapter 2: How a voice agent actually works
Anatomy of one turn
2026-09-04
Five stages run in sequence between the caller falling silent and the assistant speaking. Two of them, transport and endpointing, are excluded from the latency number Vapi displays, and Vapi says so.
Below: the paragraph from the book that builds this idea, then the diagram itself (Figure 2.1), and a recap. About a minute of reading.
Chapter 1 drew the pipeline from a distance: microphone, speech to text, model, text to speech, speaker. Figure 2.1 zooms into a single turn of that same pipeline and adds the two things the wide view leaves out: what object is actually handed from one stage to the next, and what each stage costs you.
Recap
- The idea: Five stages run in sequence between the caller falling silent and the assistant speaking.
- The picture: Figure 2.1, from chapter 2 ("How a voice agent actually works") of Vapi Made Crystal Clear.
- Go deeper: the chapter builds this step by step, with recipes and sources at the end.
This diagram is one of many in Vapi Made Crystal Clear.
Every chapter opens with the gist, draws the hard ideas, and ends with recipes and sources.
Get the book
