audio → text · 99+ languages · free
Transcribe audio to text, free.
Drop a file, get the words. Whisper large-v3 running in the browser with no account, no install and no trial to expire.
What people use this for
Interviews and research
Get a searchable transcript before you start pulling quotes. Reading is faster than scrubbing a waveform.
Meetings and calls
Turn a recording into notes you can skim. No speaker labels here, but the content is all there.
Lectures and study
A transcript is searchable in a way a two-hour recording never is — Ctrl+F beats rewinding.
Video subtitles
Export straight to .srt with timestamps already cued for your editor.
Going deeper
Starting from an MP3 specifically, need subtitles instead of a transcript, or want this running in code? And if you're weighing hosted transcription against the OpenAI API, here's the per-hour maths.
FAQ
How do I convert audio to text?+
Drop an audio or video file into the box above and click Transcribe. The file uploads, the model runs, and the text appears — typically within 5 to 15 seconds. You can copy it, download it as .txt, or export timestamped .srt subtitles. There's nothing to install and no account to create.
How accurate is automatic transcription?+
On clear single-speaker audio, Whisper large-v3 lands close to verbatim — good enough that editing is faster than typing from scratch. Accuracy degrades predictably with background noise, distance from the microphone, strong accents and people talking over one another. Nothing at any price is reliable on a phone recording from across a room.
Which languages are supported?+
99 and change, detected automatically — you don't pick one. That includes code-switching, where a speaker moves between languages mid-sentence. If you want a non-English recording turned into English text rather than transcribed in its own language, that's the translations endpoint on the API.
Is it really free, and what is the limit?+
Five files a day per IP, resetting every 24 hours, with a 25 MB cap per file. No account, no card, no expiring trial credits.
What file types work?+
mp3, wav, flac, m4a, aac, opus and ogg for audio; mp4, mov, webm and mpeg for video, where the audio track is extracted automatically. If ffmpeg can decode it, it will work here.
Can I tell who is speaking?+
Not in the free tool — it returns one continuous transcript. Speaker diarization, which labels each segment by speaker, runs on the paid API along with word-level timestamps.