T9-R3 in five minutes
LiveKit's pipeline page says voice conversations feel natural when end-to-end response latency stays under one second. That sentence is LiveKit's, not a measurement from this lab.
The same page says each stage of a cascaded pipeline adds delay, and streaming overlaps the stages.
Pipecat's overview says that in most cases the round trip happens between 500 and 800 milliseconds. That sentence is Pipecat's.
The Voice overview on docs.x.ai says the Voice APIs offer sub-second latency. The page calls them SpaceXSI Voice APIs, powered by Grok. That sentence is the page's.
This lab states no service-level agreement. The milliseconds in lab 08 are a teaching fixture.
Write each vendor sentence next to the vendor name. Do not add the three sentences into one number of your own.
Sources and go deeper
- LiveKit Agents, Pipeline types vendor last checked 2026-10
- Pipecat open source framework vendor last checked 2026-10
- Voice Overview, docs.x.ai vendor The page calls them SpaceXSI Voice APIs, powered by Grok, and separates Speech to Speech, Text to Speech, and Speech to Text. last checked 2026-10
- Python 3 tutorial vendor last checked 2026-10