All tools

MP3 to text

Turn an MP3 recording — a podcast, interview, or voice note — into searchable, quotable text.

Transcribed on your device

Local transcription benchNo file loaded
TranscribeWhisper, on-device

Speed depends on your device. The model downloads once, then works offline.

How it works

  1. Choose an MP3 file.
  2. Pick the spoken language and a model, then transcribe.
  3. Copy the transcript or download it as TXT, SRT, or VTT.

Why transcribe MP3s locally

MP3 is the default format for spoken audio: podcast episodes, interview recordings, dictation, old voice notes. Most transcription services want you to upload those files to their cloud — fine for a public podcast, less fine for an interview under embargo, a therapy-adjacent recording, or anything you would rather keep on your own machine. This tool transcribes MP3s with a Whisper model that runs inside your browser tab, so the audio never leaves your device and there is no per-minute billing.

MP3 is also the friendliest input here: the files are small, decode instantly, and the compression artifacts that MP3 introduces barely matter for speech recognition. If you have a choice of formats for the same recording, the lossless original (like WAV) can squeeze out a little more accuracy — but a decent-bitrate MP3 transcribes almost identically.

Practical notes for common MP3s

The transcript appears with timestamps as each 30-second window completes, and you can copy it or download it as plain text or subtitle files (SRT/VTT). The transcription guide explains the format choice and what accuracy to expect in more depth.

Questions

How accurate is MP3 transcription here?
Clear speech in a quiet recording transcribes well, especially with the sharper model selected. Heavy accents, crosstalk, background music, and specialist vocabulary still trip up small on-device models — treat the output as a strong first draft for anything critical.
Does MP3 compression hurt accuracy?
Barely. Speech recognition is robust to MP3 artifacts at normal bitrates (128 kbps and up). A lossless original is marginally better, but re-encoding an MP3 to WAV before transcribing gains nothing.
Can it handle a two-hour podcast?
Yes — there is no length cap, because processing happens on your device in 30-second windows. Long files just take proportionally longer, and you can stop early and keep the partial transcript.
Are different speakers identified?
No. Speaker separation requires a much larger model that is impractical in a browser tab today. You get one continuous timestamped text and attribute the speakers yourself.
Is my MP3 uploaded to a server?
No. The file is decoded and transcribed locally in your browser. The only download is the one-time model file, cached for later visits — after that the tool works offline.