Skip to content
Languages

English transcription on your own device

Transcribe English recordings with models built for English, in your browser, with nothing uploaded.

  • Nothing uploaded
  • No sign-up
  • Free
or drop it here · MP3 WAV M4A MP4 MKV MOV FLAC OGG
Language: EnglishNothing is uploaded
How it works

Three steps, all on your device

  1. Pick or record

    Choose an English recording or press Record. English is already set as the spoken language.

  2. Transcribed locally

    The Local AI engine writes the English down on your device, line by line, with timestamps.

  3. Polish and export

    Fix names and terms with Edit, then save TXT, DOCX, PDF, SRT or VTT.

Why it works

Why OmniaScribe

  • English-only models

    Each model size has an English version, usually more accurate for English than the multilingual one.

  • Private by default

    Your recording is transcribed in your browser and never sent to a server.

  • No minutes to buy

    Transcribe as much English audio as you like, free.

The best-supported language

English is where speech models are strongest, because they were trained on more English than any other language. Every model size, even Fast, is rated high for English. That makes English the language where the smaller, quicker models are most worth trying: on an older laptop or a phone, Fast often gives a transcript that needs only light editing, and Standard is enough for nearly everything else.

English-only or multilingual?

OmniaScribe offers each model size in two versions. The English-only version spends all its capacity on English, so it tends to make fewer mistakes on English speech, especially with accents, fast talkers and casual conversation. The multilingual version knows 99 languages and can translate them into English. If everything you transcribe is English, choose the English-only version on the Models screen. If you sometimes record in another language, or want translation, keep a multilingual one as well.

Accents and dialects

The models have heard a wide range of English: American, British, Irish, Australian, Indian, Nigerian, South African and many more. Strong regional accents and heavy slang still cause more mistakes than broadcast-style speech. The High accuracy model handles accents noticeably better than Fast. Spelling follows what the model heard most often, which is usually American spelling; if you need British spelling, a quick find-and-replace in your document handles “color” and “organize”.

Names, numbers and jargon

The words most often wrong in an English transcript are the ones a general model has rarely seen: people’s and company names, product codes, medical and legal terms. Numbers sometimes come out as words and sometimes as digits. After the transcript is done, search for each important name and fix it with Edit. For recurring jargon, keeping a short list of correct spellings speeds this up.

Filler words such as “um” and “uh” are usually left out, and false starts are often smoothed over. That makes the text easier to read but means it is not a strict verbatim record. For verbatim work, listen back to the lines that matter.

Recording tips for English

Speak at your normal pace and keep the microphone close. On a computer, Fast handles up to 3 h, Standard handles up to 2 h, High accuracy handles up to 1 h. On phones it's half of that. For dictation, use Record: a live draft appears as you talk, and the final text is made when you stop.

Live dictation in English

Because even the smaller models do well in English, the live draft you see while recording is unusually good, and it is often close to the final text. That makes English the easiest language for dictating emails and notes: speak, watch the words appear, and press Stop when you are done.

Specs

Formats and limits

Audio
MP3, WAV, M4A, AAC, FLAC, OGG, OPUS
Video
MP4, MOV, MKV, WEBM
Export
TXT, PDF, DOCX, SRT, VTT
Languages
99, or detect automatically
Longest file per model
ModelComputerPhone
Fast3 h1 h 30 min
Standard2 h1 h
High accuracy1 h30 min

Files that would take more than about 45 min on your device ask before they start.

Expected accuracy in English

From published test results for each model size, on clear read speech. Your own recordings may do better or worse.

  • FastHigh
  • StandardHigh
  • High accuracyHigh
FAQ

Questions, answered.

Something else on your mind? Email us and a person replies.

Which model is best for English?

The English-only version of the size your device runs well. Standard suits most devices; High accuracy is best for accents and difficult audio.

Is my English recording uploaded?

No. It is transcribed in your browser on your device.

Does it handle British, Indian or Australian accents?

Yes. Strong accents cause more mistakes than standard speech, and High accuracy handles them best.

Does it use American or British spelling?

Usually American spelling. Change it with find-and-replace in your document if you need British spelling.

Are “um” and “uh” included?

Usually not. The transcript reads cleanly but is not strictly verbatim.

Can I translate other languages into English?

Yes, with a multilingual model and Translate to English switched on.

Guides

OmniScribe for Android

Transcribe anywhere.
No signal needed.

Install the free Android app and turn recordings into text on your phone, even with the screen off. Your audio never leaves the phone.

  • Works offline after a one-time model download
  • Translates offline, too
Get it on Google Playfree with ads · optional Pro
The OmniScribe home screen