Cost and latency

Cost

Every session shows what it cost, broken down by component - the transcriber, the model, the voice, and the platform fee - rather than as a single number you have to trust.

A component you run on your own key does not appear as a charge, because Glytos is not buying it for you.

How the number is arrived at is on Pricing and credits.

Latency

A voice call feels good or bad based on the gap between the caller finishing and the agent starting. Each session reports where that gap went:

StageWhat it measures
TranscriberTime to turn speech into text.
ModelTime to the first token of the reply.
VoiceTime to the first audio of the reply.
EndpointingTime spent deciding the caller had finished.
TurnThe whole gap, end to end.

Knowing which stage is slow is what tells you what to change - a faster model, a different voice provider, or turn-taking that is waiting too long. Guessing at this without the breakdown usually means changing the wrong one.

Cost and latency · Glytos Docs