Transcribe Audio to Text
Turn any recording into accurate text. Upload audio, get a clean transcript back in minutes, with automatic language detection across 90+ languages.
Transcribe Audio to Text
Upload a recording and get an accurate, readable transcript back.
Typing out an interview by hand takes roughly four hours per hour of tape. This does it while you get on with something else. Drop in a recording of a meeting, interview, lecture, podcast, or voice memo and get clean text you can search, quote, edit, or paste straight into your notes. Language is detected automatically, so you do not have to tell it what you are working with.
How to transcribe audio to text
- Upload your audio file.
- Leave the language on auto, or set it if you already know.
- Optionally add names or technical terms so they come back spelled correctly.
- Run it and copy your transcript.
Features
- 90+ languages, detected automatically. You do not need to specify the language unless you want to force one.
- Spelling hints. Pass in names, places, product names, or jargon up front and they come back spelled properly instead of phonetically guessed.
- Handles real recordings. Built for the messy end of the spectrum: accents, background noise, phone audio, and people talking over each other.
- Nothing to install. Runs in the browser, so it works the same on a Mac, a PC, or a Chromebook.
Getting names and jargon right
The single biggest cause of a transcript needing cleanup is proper nouns. Speech models guess phonetically, so a surname, a drug name, or an internal product codename comes back spelled the way it sounded.
The optional hint field fixes this. List the terms you expect to hear before you run it, and they will be spelled the way you wrote them. For a company all-hands that might be the exec team’s names plus your product names; for a medical or legal recording it is worth pasting in the specialist vocabulary. It takes ten seconds and removes most of the find-and-replace afterwards.
FAQ
How do I transcribe audio to text?
Upload the recording and the transcript is generated for you. There is no manual typing and no need to play the audio back in real time. Most files finish in a fraction of their runtime.
What audio formats can I upload?
Common formats including MP3, M4A, WAV, and the audio from most recording apps. Files can be up to 25MB. If your recording is longer than that, split it or compress it to a lower bitrate first.
Can I transcribe a video?
Not directly here, since this takes an audio file. Run your video through Extract Audio from Video first, then transcribe the audio it gives you.
How accurate is it?
Accuracy is high on clear speech and drops with background noise, heavy crosstalk, and strong accents. Expect to skim and fix a handful of words rather than retype anything. Adding spelling hints for names and jargon makes the biggest single difference.
Does it work in languages other than English?
Yes, across 90+ languages, detected automatically. If a recording switches between languages, setting the primary language explicitly usually gives a steadier result than leaving it on auto.
Can I transcribe a meeting or an interview?
Yes, that is the common use. One thing to know: the transcript comes back as continuous text rather than split by speaker, so for a multi-person recording you will need to mark up who said what yourself.