Convert audio to text — free and unlimited

Drop in almost any audio file — MP3, WAV, M4A, FLAC, OGG — and Whisper turns it into an editable transcript with timestamps. Switch on speaker labels to see who said what, for up to 8 speakers, then download it as SRT or VTT subtitles, a Word document or plain text. There is no minute cap, no account, and no queue. Your files are never uploaded — everything runs on your device.

Most transcription sites limit the free tier because every minute costs them server money. This one runs the speech model inside your own browser, on your own hardware, so there is nothing to meter. The engine is open source, and after the first model download it keeps working with the connection off.

Checking what this browser supports…

How to transcribe audio

  1. Pick your audio file

    Drag it into the box above or click Choose file. Common audio formats are supported directly — no conversion needed first.

  2. Wait for the one-time model download

    On first use the speech model (about 80 MB) downloads into your browser cache. It stays cached, so later transcriptions start immediately.

  3. Watch the transcript stream in

    Long recordings are processed in chunks, so text starts appearing before the file is finished. Progress and a time estimate are shown throughout.

  4. Edit and export

    Click any segment to fix a word, then download the result as TXT, SRT, VTT, JSON or a Word document — or copy the plain text.

Why this converter is different

Actually unlimited

No minute caps, no monthly quota, no premium tier. Because inference runs on your device, a three-hour file costs us the same as a three-second one: nothing.

Private by design

The audio is decoded and transcribed locally. Privacy here is not a policy promise — uploading your file is simply not part of the design.

No signup, no nagging

Open the page, drop a file, leave with your transcript. No account, no email, no trial countdown.

Frequently asked questions

Is it really free without limits?
Yes — YourDevice transcription is free with no account, no watermark, no per-file cap and no limit on how long a recording can be. Whisper runs in your browser on your own hardware, so there is no per-minute server cost for us to pass on; the site is paid for by ads rather than by metering the tool.
Which audio formats work?
Anything your browser can decode: MP3, WAV, M4A/AAC, FLAC and OGG cover almost everything. Video files work too — YourDevice extracts the audio track in the browser, so an MP4 needs no converting first. The finished transcript downloads as plain text, SRT or VTT subtitles, or a Word document.
How accurate is the transcription?
YourDevice transcribes with Whisper, OpenAI’s open-source speech-recognition model and one of the most widely used available. Accuracy depends on audio quality, accents and background noise, so we won’t quote a single percentage — a clean interview and a recording made across a noisy room are not the same job. Run a real file through it and judge the result; your recording is never uploaded either way.
Which languages are supported?
Forty, listed by name — every European language plus the major world languages, curated from the roughly 99 Whisper knows. You pick the spoken language yourself before transcribing. There is no auto-detect, deliberately: the underlying library has not implemented language detection, so an “Auto” option would quietly transcribe German audio as English rather than admit it did not know. There is also a translate option that outputs English regardless of the source language.
Can it tell who is speaking?
Yes, for up to 8 speakers. Switch on Identify speakers and each line is labelled Speaker 1, Speaker 2 and so on; rename them and the names carry into the SRT, VTT, Word and text exports. The labels come from NVIDIA's Nemotron 3 Diarization model, a 120 MB download the first time, which runs in your browser after the transcript is done.
Does it work offline?
Yes, once you have used it. Your first visit installs a small service worker, so reopening the page with the connection off serves it from your browser's cache, and the transcription then runs on the Whisper model already stored there — we test that by transcribing a file with the network disabled. Two things still need a connection: a page you have never opened, because nothing is downloaded ahead of time, and a model you have not downloaded yet.
Where does my file go?
Nowhere. YourDevice reads, decodes and transcribes the recording inside your own browser tab. Your files are never uploaded to any server — ours or anyone else’s — and nothing is kept once you close the tab.
How can I check that my audio file is not being uploaded?
Open your browser's developer tools, switch to the Network tab, and start a transcription with it open. On the first run you will see the Whisper model download once — about 80 MB for the standard model, less or more if you pick a different one — and after that it is served from your browser's cache. What you will not see, on any run, is a request carrying your audio file. Then sort that list by size: the largest thing the page sends is a few kilobytes of analytics, nowhere near the size of a recording. This page loads analytics and advertising like most websites do, which is why the claim is scoped precisely: your files are never uploaded.