AI glossary · Voice
Text to speech
In one sentence
Text to speech (TTS) is the technology that turns written text into spoken audio. It's the component, usually a model reached through an API, that gives a voice to a phone agent, a screen reader or any audio generated from text.
Also known as: TTS · text-to-speech conversion · AI voice · synthetic voice
What is text to speech?
Text to speech, shortened to TTS, is a machine’s mouth. You give it text and it returns audio of a voice reading it. A few years ago synthetic voices sounded flat and robotic. Today’s, generated by neural networks, have intonation, pauses and rhythm very close to a person’s.
Put it to work
A voice agent that picks up the phone
From €149/mo + VAT
It answers calls 24/7, books appointments and logs every call in the CRM.
See the voice agentTwo neighbouring terms:
- Text to speech: the specific component or service that turns text into audio.
- Speech synthesis: the technical discipline that makes it possible.
Its opposite is speech to text, which turns voice into text.
Related to this term
UTIO Voice
- 500 minutes and a number included
- Natural conversation in 26 languages
- Introduces itself as a virtual assistant
- 14-day free trial, no card needed
From €149/mo + VAT
See the voice agentHow it works
- Normalise the text. Expand abbreviations, numbers and dates: “£15” becomes “fifteen pounds”.
- Work out pronunciation and prosody. Which sounds to make and with what intonation: a question isn’t said the same way as a statement.
- Generate the audio. A neural model produces the sound wave in the chosen voice.
What you can usually adjust:
| Setting | Example |
|---|---|
| Voice | Male, female, younger, deeper |
| Language and accent | British English, Spanish, French |
| Speed and pauses | Slower when reading out a phone number |
| Pronunciation | How to say a brand name or acronym |
In a voicebot, TTS works in streaming mode: it starts speaking with the first words of the reply instead of waiting for the full sentence. That’s key to keeping latency down.
Example
A regional bus company in Yorkshire has to announce timetable changes by phone and over station tannoys. It used to record each announcement with a voice artist in a studio. Now the team writes the announcement and a TTS service produces the audio in minutes, always in the same voice. To get local place names right, they add a few pronunciation rules of their own.
Text to speech at UTIO
UTIO’s voice agent uses TTS to speak on every call: it introduces itself as a virtual assistant and converses in 26 languages. The voice is included in the minute, along with the model, transcription and telephony; voice synthesis comes from US providers under contractual safeguards.
If you’re a developer, UTIO’s AI API offers text to speech: natural audio from text, in several languages and voices, with the same key as transcription, OCR or embeddings.
Frequently asked questions
What is the difference between text to speech and speech synthesis?
Speech synthesis is the discipline: how artificial speech is produced. Text to speech is its most common application, with text as the input. A “TTS” service is packaged speech synthesis.
Can I choose the voice and accent?
Yes. TTS services offer catalogues of voices by language and accent, and let you adjust speed and pauses. Some accept markup in the text, such as SSML, to say how a date or an acronym should be read.
How does TTS read numbers, acronyms and addresses?
That's where it slips most. Before synthesis, the text is “normalised”: “St” becomes “Street” or “Saint”, “3/10” becomes a date. Test with your own data and, where needed, spell out in the text how a word should sound.
Related terms
- Speech synthesis
Speech synthesis is the artificial production of human speech by computer. It's the discipline behind text to speech, voice cloning and talking agents, and today it relies mainly on neural networks trained on real recordings.
- Speech to text
Speech to text (STT) is the technology that turns spoken audio into written text. It's the software component, usually a model reached through an API, that does the listening in a voicebot, a dictation tool or a transcription service, either live or on recordings.
- Voicebot
A voicebot is a program that holds spoken conversations: it listens to what a person says, works out what they want and answers out loud. It can run on a phone line, a website, an app or a smart speaker, and with generative AI it handles open conversation, not just menus.
- Latency
Latency is the time between making a request and starting to receive the response. In a voice agent, it's the silence from the moment the caller stops talking to the moment they hear the agent's first word, and it decides whether the call sounds natural.
Updated on