Latency
Latency is the time from when a process is initiated until it actually begins to run. In a conversation with an agent or an API call, it is the time elapsed before a response is received.
On this page
Source content checked . English translation published 2026-10-27.
In business automation
In a conversation with a voice agent or an API call, the time before a response starts affects the experience more than the length of the response itself. Latency is the time from the start of an action until execution actually begins. In the guide, the term appears alongside voice agents and timeouts, without making any claims about the speed of a particular tool.
Examples
These examples illustrate a possible workflow. Adapt the connections and rules to your own systems.
Waiting for a phone agent to start speaking
A team builds a phone agent. The customer asks a question, and the system sends the text to a model and then to speech synthesis. If latency is high, the customer repeats the question. The team measures the time until speech begins without setting an untested numeric target.
Relevant tools
- LiveKit
- A framework for building real-time agents. Latency depends on the entire conversation chain. This does not include a measurement of the tool.
What to check
- Confusing latency with a timeout: latency is how long it takes for a response to start, while a timeout is when waiting stops.