MEDIA TRANSCRIBER

Audio and video to text.
Free, never uploaded.

Pick a recording, a voice memo or a video and get a timestamped transcript. Free, no account, no catch: the speech recognition runs in your browser, so the file stays on your computer or phone.

HOW IT WORKS

Three steps to a transcript

  1. 01Pick a file

    Choose an audio or video file from your device, or drop it onto the page. MP3, MP4, MOV, MKV, WebM, M4A, WAV, OGG and FLAC all work.

  2. 02Let your device listen

    The first time, your browser downloads a speech model and keeps it for next time. The text appears as it goes, a section at a time.

  3. 03Copy or download

    Copy the transcript, or save it as plain text, or as SRT or VTT subtitles with timestamps. It doesn't label different speakers.

WATCH IT

Watch the media transcriber at work

Pick an MP3, MP4 or other media file and watch the text appear as it goes. This part of our walkthrough starts at 11:38. It loads from YouTube only when you press play. Watch on YouTube

Free, with no signup and no watermark. How long a file can be depends on your device's memory. Nothing is uploaded, so your file stays on your device.

QUESTIONS

Frequently asked

Is the media transcriber really free? What's the catch?

Yes, and there isn't a catch. No account, no ads, no paid tier and no watermark. How long a file can be depends on your device's memory. It's a one-person project by Gerald Ho (Axecute), and the only money is an optional PayPal tip.

Is my file uploaded anywhere?

No. The file is read and transcribed inside your browser. Your browser downloads the speech model from Hugging Face and some supporting code from jsDelivr, and the page loads Google Analytics. None of these sends anything from your file. Your audio and video never leave your device, and shinmuscat never sees them.

Which file formats work?

Common audio and video formats: MP3, MP4, MOV, MKV, WebM, M4A, AAC, WAV, OGG, Opus and FLAC. Video files work too, and only the sound is used. If a file will not open, try Chrome or Edge, or convert it to MP3 or MP4 first.

How does it transcribe without a server?

The first time you use it, your browser downloads a speech-recognition model (Whisper) and keeps it for next time. Your device then runs the model itself, using the graphics card where the browser supports it and the processor otherwise.

How long does it take?

It depends on your device. On a recent laptop with a supported graphics card it usually runs much faster than the recording plays. On an older computer or a phone it can take as long as the recording or longer. The transcript fills in as it goes, so you can read the start before it finishes.

Does it work on my phone?

Yes, in a recent version of Safari, Chrome or another mobile browser. Keep the tab open and the screen on until it finishes, because phones pause pages in the background. Long files are faster on a computer.

Which languages does it understand?

Whisper detects the spoken language on its own and handles many languages, with English the most accurate. It does not label different speakers.

RELATED

More guides