Text-to-speech (TTS)
Text-to-speech turns written text into spoken audio, and is the voice a caller hears from a voice agent.
Modern neural TTS produces natural-sounding speech with realistic pacing and intonation, in many languages and voices. For live calls it streams: the first words start playing while the rest of the sentence is still being generated, which keeps the reply fast.
The choice of voice and provider affects how human the agent sounds, how it handles names, numbers and mixed languages, and how quickly the first audio arrives.
