What Is Speech to Text? A Plain Explanation

What is speech to text? It is software that converts spoken language in audio into written text. The program cuts the sound into short frames, matches them to speech sounds and builds the most likely words. The same idea goes by the names speech recognition and ASR.

Transcribe a recording

Send your voice message, audio or video to @potok_now_bot on Telegram and the text comes back in the chat.

Open @potok_now_bot in Telegram

Files up to 2 GB · export to TXT, MD, DOCX, PDF, SRT and VTT · speaker labels and timestamps · about 30 languages

Key facts

Modern systems use neural networks trained on large collections of recordings paired with correct text, not hand-written rules.
The usual quality measure is word error rate, the share of words that are replaced, missed or added compared with a correct transcript.
Accents, jargon, crosstalk and noise raise the error rate, and so does a language the system saw little of in training.
The raw output has no sense of who spoke or where the sentences end, so extra steps add speaker labels, punctuation and timestamps.
Speech to text turns sound into words in the same language; it does not turn one language into another.

What it looks like

// input
talk.m4a  (spoken: "send the invoice on friday")

// pipeline
audio -> frames -> speech sounds -> likely words -> text

// output
send the invoice on friday

See the full list of languages, the tools and the glossary.

Frequently asked questions

Is speech to text the same as dictation?

Dictation is a use of it: you talk and the text appears in a document. Transcription of a recording is another use of the same technology.

Is it the same as speech recognition?

Yes, the terms are used interchangeably. ASR, short for automatic speech recognition, is the technical name.

How accurate is it?

It depends on the audio, the voice and the language. Published word error rate figures show what to expect, and a clear recording helps most.

How do I convert a recording to text with Potok?

Open the bot, send the voice message, audio or video, or paste a YouTube link, and the text returns in the chat or as a file.

Can I export the result?

Yes. Exports include TXT, MD, DOCX, PDF, SRT and VTT. Short results arrive as text in the chat and long ones as a file.

Sources

Each source was opened by hand on the date shown. Pages change, so open the source if a detail matters to you.

Speech recognition (Wikipedia), Wikipedia · checked 2026-10-18
Word error rate (Wikipedia), Wikipedia · checked 2026-10-18

Try it on your own recording