MP3 to text, without the upload

MP3 is where the world’s spoken audio ends up: podcast episodes, recorded interviews, lecture archives, old dictaphone files. This page turns those MP3s into searchable, editable text — right here in the browser tab, with nothing sent anywhere.

A one-hour podcast episode would take a while just to upload on most connections. Here there is no upload step at all: the file is read from disk and transcribed on the spot, with the transcript streaming in as it goes. Your files are never uploaded — everything runs on your device.

Checking what this browser supports…

How to convert an MP3 to text

  1. Add your MP3

    Drag the file into the tool above. Long episodes are fine — length only costs your machine’s time, not money.

  2. Let it transcribe in chunks

    The recording is processed in 30-second windows, so the beginning of a long interview is readable while the rest is still being worked through.

  3. Clean up and export

    Fix names and jargon with a click, search within the transcript, then export TXT for notes, SRT/VTT for captions, or JSON for further processing.

Made for long-form audio

Podcast-length friendly

No per-minute billing means a full season backlog is just a matter of letting it run. Progress and an ETA keep you informed on longer files.

Compressed audio, handled

MP3’s compression is no obstacle — the decoder your browser already ships handles bitrates from voice memos to studio masters.

Interviews stay confidential

Unpublished interviews and embargoed material never leave the machine. There is no server copy to worry about, because there is no server involved.

MP3 transcription questions

How long can the MP3 be?
There is no enforced limit. Practical length depends on your device’s memory and patience — hour-long episodes are routine, and the transcript streams in as it processes so you can start reading early.
Does low bitrate hurt accuracy?
Heavily compressed or noisy recordings transcribe less accurately than clean ones — that is true of every transcription system. Speech stays intelligible to the model at bitrates far below typical podcast quality.
Can I get timestamps for each part?
Yes. Every segment carries start and end times, visible in the transcript view and preserved in SRT, VTT, and JSON exports.
What engine does the transcription?
It is powered by OpenAI’s open-source Whisper model, running in your browser through WebGPU or WebAssembly. Our engine wrapping it is open source as well.
Why is there no upload progress bar?
Because your file is never uploaded — it is read straight from your disk into the browser’s memory. The only download is the model itself, once, on your first visit.
How can I check that my MP3 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 MP3. Then sort that list by size: the largest thing the page sends is a few kilobytes of analytics, nowhere near the size of an MP3. This page loads analytics and advertising like most websites do, which is why the claim is scoped precisely: your files are never uploaded.