What is text to speech (TTS)?

The system that turns a voice agent's written reply into the audio a caller actually hears.

An admin approves every new account by hand. Nothing is created until then. We reply by email; no newsletter, no sequence.

app.salescrew.io/inbox
The unified reply inbox with classified threads

Text to speech (TTS)Text to speech (TTS) converts written text into spoken audio, and in a voice agent it produces every word the caller hears. Voice choice, prosody and streaming speed set how human the agent sounds and how much of the per-minute cost it consumes.

Also known as: TTS, speech synthesis

Why it matters

A language model produces text. On a phone call, that text has to become sound before it means anything to the caller, and that conversion is entirely TTS's job. The voice a caller hears, its accent, tone, pacing and how naturally it handles punctuation and numbers, all comes from the TTS engine, not the language model that decided what to say.

Two things about TTS decide whether a voice agent feels usable: how quickly it starts producing audio once it has text to speak, since a slow TTS engine adds directly to voice latency, and how expensive each voice is per character or per minute. Premium, highly natural voices generally cost more to run than flatter, faster ones, and that trade-off shows up directly in the per- minute bill for a voice agent.

What TTS does in a call

  1. 1

    The model produces text

    The language model's reply is generated as written words, not audio.

  2. 2

    TTS converts it to audio

    The text is synthesised into speech in the configured voice.

  3. 3

    Audio streams to the caller

    Well-built systems start streaming the first words before the full sentence finishes generating, to cut latency.

  4. 4

    The voice, pace and tone are set upfront

    Choices made when configuring the agent, not decided per call.

The mistake to watch for

Picking the most expressive voice without checking its price per character. Premium voices can be the largest line in the per-minute cost.

Questions

What is the difference between text to speech and speech to text?
They run in opposite directions. Text to speech converts written words into spoken audio, producing what the caller hears from the agent. Speech to text converts the caller's spoken audio into written words the language model can read. A voice agent needs both.
Does the choice of TTS voice affect cost?
Yes, often significantly. TTS providers typically price by character or by minute of generated audio, and more expressive, natural-sounding voices are commonly priced higher than simpler ones, which makes voice choice a real line item in a voice agent's per-minute cost.
Can TTS handle numbers, dates and prices correctly?
Good TTS engines normalise numbers, dates, currency and abbreviations into how a person would actually say them, rather than reading digits or symbols literally. This is worth testing directly with the specific content a business's agent will actually speak, like prices or addresses.