Free audio to text converter
Transcribe Audio to Text Free
Drop in an audio file and get a free, editable transcript in your browser. MP3, WAV, M4A, MP4 and more. No upload and no sign-up.
How to transcribe audio to text
- Drop your file in the box above, choose it from your device, or press Record from mic.
- Pick Fast for a quick draft or Accurate for cleaner text. Leave the language on Detect unless the speaker uses a language other than English.
- Press Transcribe. The first run downloads the speech model. Text appears as it's recognized.
- Fix any names or terms the model missed, then copy it, download a .txt file, or download .srt captions.
Want to know how long the transcript runs when you read it aloud? Paste it into the words to minutes calculator.
A private audio to text converter
Any audio file you drop here is transcribed in the browser tab. Script Timer never receives the recording, and nothing is stored on a server. The only download is the speech model, once, from the Hugging Face hub. Your browser keeps it for the next file.
That matters for interviews, client calls, legal notes and anything else you'd rather not hand to a stranger's server. It also means there's no queue. Your file starts the moment you press the button.
Convert MP3, MP4, M4A and WAV to text
| Format | Where it usually comes from |
|---|---|
| MP3 | Podcasts, music-app exports, most recorders |
| M4A and AAC | iPhone Voice Memos, many Android recorders |
| WAV | Studio recorders, Zoom local audio, DAW exports |
| MP4 and MOV | Video files. The audio track is pulled out in your browser |
| OGG, OPUS, WEBM, FLAC | Browser recordings, voice notes, lossless archives |
If one browser can't read a file, try Chrome or Edge on a computer. They decode the widest range of formats.
How to transcribe voice memos
iPhone Voice Memos save as M4A. Share the memo to Files, drop it in above, and press Transcribe. The full walkthrough is in how to transcribe voice memos. Android recordings work the same way once you save a file.
Fast or Accurate
Fast uses the smallest Whisper model. It downloads quickly and works on phones, which makes it good for a rough draft or a clear single speaker.
Accurate uses a larger model. It's slower, but it handles accents, cross-talk and background noise better. On a laptop it's the better default.
Neither one labels who's speaking. If you need speaker names, add them as you edit.
What to do with an audio transcript
Treat the transcript as a first draft of the recording. Paste it into the words to minutes calculator, cut the filler, and read the clean version on the free teleprompter before the next take.
To hear that draft out loud, Script Timer reads it back in 100+ AI voices, so the clunky lines show up before an audience does.
More free audio converters
Audio to Text FAQ
Is my audio file uploaded anywhere?
No. The audio file is read and transcribed inside your browser. Nothing is sent to Script Timer. The speech model downloads once from the Hugging Face model hub and your browser caches it.
Is audio transcription free?
Yes. Transcribing an audio file needs no sign-up and adds no watermark. One file can run up to 60 minutes, which is as long as the browser can handle in a single pass.
What speech model transcribes the audio?
OpenAI's open-source Whisper model, run in the browser with transformers.js, the same approach the open-source whisper-web project uses. Fast uses Whisper Tiny. Accurate uses Whisper Base.
Why is the first audio file slower?
The first audio file waits while your browser downloads the speech model. Later files use the cached model and start right away. Speed after that depends on your computer or phone.
How accurate is the transcript?
Clear audio with one speaker comes out close to clean. Noise, music, overlapping voices and heavy accents add mistakes. Accurate mode helps. Always read it through before you publish it.
Can I get captions with timestamps?
Yes. After a run, press Download .srt for a captions file you can upload to YouTube, Vimeo or most video editors, or download a plain text file with a timestamp on each line.
Hear it before your audience does
Paste your script into Script Timer to time it, grade the delivery and hear it in 100+ AI voices.
In-browser speech recognition follows whisper-web (MIT license).