ONLINEAGENT_OPS 2026.Q3 HOME HUBSMODELSBUILDINGCOMPANIESFILMMAKINGSECURITYLIVING WITH ITSKILLSALL HUBS, ONE TERMINAL → ARTICLES CRAFT RECORD BLOG TOOLS SKILLS SEARCH /
HOME›AI MEANINGS›WHAT IS SPEECH-TO-TEXT?
AI MEANINGS · MEANINGBEGINNER

What is speech-to-text? AI that writes down what you say

Speech-to-text means software that listens to speech and writes it down, as in dictation and captions. Also called speech recognition or ASR.

READ5 min
WORDS977
SECTIONS6
TYPEMEANING
CHECKED5 OCT 26

Speech-to-text means software that listens to spoken words and writes them down as text. It is what powers dictation on your phone, automatic captions on videos and meeting transcripts. It is also called speech recognition, or ASR (automatic speech recognition).

◈ IN PLAIN TERMS

Think of a typist sitting in on every conversation, writing down each word as it is spoken. Speech-to-text is that typist, built into software: it never gets tired, but it can still mishear a name, an accent or a word it has never come across.

In a sentence: “I record my lectures and let speech-to-text turn them into notes, then fix the names it gets wrong.”

What does speech-to-text mean?

The plain point: sound goes in, written words come out.

IBM, the computer company, defines it in one sentence: “Speech recognition, also known as automatic speech recognition (ASR), computer speech recognition or speech-to-text, is a capability that enables a program to process human speech into a written format.”1 Hugging Face, a large hub where AI models are shared, calls it “the task of transcribing a given audio to text”.2

The better systems today are built with machine learning: they learn from examples of speech instead of following rules a person wrote.1

A simple example

The plain point: you probably use it already without thinking about it.

When you say “set a timer for ten minutes” to a phone, speech-to-text turns your voice into words before the assistant works out what you want. Hugging Face lists voice assistants and automatic captions for videos among the main uses; captions also make videos accessible to people who cannot hear the sound.3 IBM adds doctors dictating patient notes and call centres transcribing calls.4

How accurate is it?

The plain point: best with clear speech in a quiet room; worse with accents, noise and unusual words.

IBM names the usual troublemakers: “pronunciation, accent, pitch, volume, and background noise”, plus specialist jargon.1 Accuracy can also differ between groups of people. A 2020 study tested speech-to-text systems from Amazon, Apple, Google, IBM and Microsoft and found they got almost twice as many words wrong for Black speakers as for white speakers: an average error rate of 0.35 against 0.19.5 That study is several years old, and the systems have changed since; the lesson is that the people a system heard least during training tend to be misheard most. AI bias meaning explains why.

TAKEAWAY

Before you rely on a transcript, check the names, numbers and technical words. Those are what speech-to-text gets wrong most often, and the mistakes look perfectly confident on the page.

Often confused with

TermWhat it doesHow it differs
Voice recognitionDoes: works out who is speaking.Differs: speech recognition works out what was said. IBM says voice recognition “just seeks to identify an individual user’s voice”.1
Text-to-speechDoes: reads written text out loud.Differs: the opposite direction: text in, voice out.
Voice cloningDoes: copies a particular person’s voice.Differs: it makes speech; speech-to-text only listens.
Machine translationDoes: turns one language into another.Differs: live-translation apps often do speech-to-text first, then translate the words.

Questions people ask

What is the difference between speech recognition and voice recognition?

Speech recognition writes down what was said. Voice recognition identifies the person speaking, for example to unlock a device or check a caller’s identity, as in the IBM definition above.

Is speech-to-text the same as text-to-speech?

No, they are opposites. Speech-to-text listens and writes; text-to-speech takes writing and speaks it aloud. Many voice assistants use both, one after the other.

How accurate is speech-to-text?

It works best with clear speech in a quiet room. IBM lists accent, pitch, volume, background noise and specialist jargon among the things that make it less reliable, and the 2020 study above found it made almost twice as many mistakes for one group of speakers as for another.

Learn more on this site

QUICK CHECK · 3 QUESTIONS

Pick an answer to see if you got it. Nothing is recorded.

1. What does speech-to-text do?

  • It reads written text out loud
  • It works out who is speaking
  • It writes spoken words down as text
Show the answer

Answer: It writes spoken words down as text. The page says speech-to-text listens to spoken words and writes them down: sound goes in, written words come out.

2. Before you rely on a transcript, what should you check first?

  • Only the first and last sentences
  • Names, numbers and technical words
  • The font and layout of the text
Show the answer

Answer: Names, numbers and technical words. The page says names, numbers and technical words are what speech-to-text gets wrong most often, and the mistakes look confident.

3. How is speech recognition different from voice recognition?

  • Speech recognition finds what was said; voice recognition finds who said it
  • They are two names for exactly the same thing
  • Voice recognition writes words down; speech recognition reads them aloud
Show the answer

Answer: Speech recognition finds what was said; voice recognition finds who said it. The page says speech recognition writes down what was said, while voice recognition identifies the person speaking.

Related terms: LoRA · Negative prompt · Seed (in AI images) · Style reference · Text-to-image · Text-to-speech (TTS)

SOURCES
  1. 1IBM, What is speech recognition?, 28 Sep 2021, read at source 5 Oct 2026“Speech recognition, also known as automatic speech recognition (ASR), computer speech recognition or speech-to-text, is a capability that enables a program to process human speech into a written format.” ↑a“Many speech recognition applications and devices are available, but the more advanced solutions use artificial intelligence (AI) and machine learning.” ↑baccuracy is affected by “pronunciation, accent, pitch, volume, and background noise”. ↑c“Speech recognition focuses on the translation of speech from a verbal format to a text one whereas voice recognition just seeks to identify an individual user’s voice.” ↑d
  2. 2Hugging Face, Automatic Speech Recognition, read at source 5 Oct 2026“Automatic Speech Recognition (ASR), also known as Speech to Text (STT), is the task of transcribing a given audio to text.” ↑
  3. 3Hugging Face, Automatic Speech Recognition, read at source 5 Oct 2026, which lists virtual speech assistants and caption generation for live-streamed or recorded videos as use cases, including to improve accessibility.↑
  4. 4IBM, What is speech recognition?, 28 Sep 2021, read at source 5 Oct 2026, which lists uses in cars, healthcare dictation, call centres and security.↑
  5. 5Koenecke et al., Racial disparities in automated speech recognition, PNAS, 23 Mar 2020, read at source 5 Oct 2026“all five ASR systems exhibited substantial racial disparities, with an average word error rate (WER) of 0.35 for black speakers compared with 0.19 for white speakers.” ↑

Koenecke et al., Racial disparities in automated speech recognition, Proceedings of the National Academy of Sciences, 23 Mar 2020, read at source 5 Oct 2026.

The comparison on this page (a tireless typist), the timer example, the point about live-translation apps and the advice on checking transcripts are this site’s own ways of explaining, not claims from these sources.

Free printable AI cheat sheets, and AI news every week.

Sixteen one-page AI guides and checklists in one free PDF, no sign-up. And if you want to keep up, the newsletter covers what changed in AI each week, in plain English, with the sources.

SUBSCRIBE FREE ↗ GET THE CHEAT SHEET PACK
aipromptgeneer://$ about this siteAI PROMPT GENEERAI explained in plain English. Every claim links to its source, with the date it was checked.NEWSLETTERWhat changed in AI each week, in plain English.SUBSCRIBE FREE ↗AI PROMPT GENEER · MACHINE ASSISTANCE DISCLOSEDEVERY ARTICLE CARRIES ITS OWN CHECKED DATE