Skip to content
Technology

Speech-to-Text (STT)

AI technology that transcribes spoken audio into written text. Handles accents, background noise, multiple speakers, and technical jargon. Also called Automatic Speech Recognition (ASR).

Why it matters

STT turns talking into typed text, which saves enormous time whenever writing by hand is slow or impractical. It matters for anyone who dictates notes, captions videos, or transcribes interviews and meetings. Good STT handles accents, background noise, and multiple speakers, making spoken words searchable and shareable. It's also a big accessibility win, giving deaf and hard-of-hearing people real-time captions for calls and videos.

A concrete example

A journalist records a 40-minute interview, then runs it through an STT tool. Minutes later she has a full transcript that even labels who said what, despite some traffic noise and the subject's strong accent. Instead of replaying the audio for hours to find quotes, she searches the text for a keyword and jumps straight to the moment she needs. What was tedious manual work is now a quick, searchable document.

How to use it

Recording quality determines accuracy more than the model does: one speaker per microphone, minimal background noise, and no talking over each other will improve a transcript more than switching provider. Give the system a list of names, jargon and acronyms in advance if it supports one. Always read a transcript before publishing it as an accessibility artefact — an unreviewed transcript with errors is worse than none.

The common mistake

Trusting speaker labels in a multi-person recording. Diarisation regularly swaps who said what, which turns a transcript into a confidently wrong record of who committed to which action.

Related terms

Put Speech-to-Text (STT) into practice

Access 750+ AI models and 60+ tools through Vincony — start free with 100 credits.