Convert audio to text
Transcribe interviews, lectures, voice notes and meetings. The Whisper speech model runs on your device, so the recording never leaves it, and after a one-time setup it works even offline.
Free forever · Runs on your device · 0 bytes uploaded
Speed depends on your device; big files on phones take longer. Slow or not working? Report it on GitHub
How to audio to Text
- Choose an audio or video file (MP3, M4A, WAV, OGG, MP4 or WebM).
- Pick the spoken language, or leave it on auto-detect, and choose text or subtitles.
- Start it. The first time, the speech model downloads once; then read the text and download it.
Also good for
- transcribe audio to text free
- speech to text online
- voice recording to text
- transcribe interview or lecture
- private transcription without upload
Private by design — and you can check
Your files are never uploaded, because there is nowhere to upload them to.
- No upload step existsFiles are read and written inside this browser tab. There is no server that could receive them.
- The browser enforces itThis page’s Content Security Policy only allows connections back to this site, which serves nothing but the app itself.
- See for yourselfOpen developer tools, watch the Network tab and run a job: no request carries your document.
- Open sourceEvery line is on GitHub, built on qpdf, pdf.js & pdf-lib.
Frequently asked questions
Are my files uploaded anywhere?
No. Fizzdoc does the work inside your browser tab, using open-source engines compiled to WebAssembly and JavaScript. Your file is read from your device, processed locally, and the result is saved straight back to your device. You can check this yourself: open your browser’s developer tools, go to the Network tab, and run a job — no request carries your file.
How accurate is it?
Fizzdoc uses Whisper tiny, a small version of OpenAI’s open speech model that runs fully in your browser. Clear English speech comes out well; noisy recordings, strong accents and other languages are rougher, so read the text through before you rely on it.
Why is the first run slower?
The first time, your browser downloads the speech model (about 58 MB) and keeps it, so later runs start in seconds, even offline. Transcribing takes a while for long recordings: speed depends on your device, and a laptop is much faster than a phone.
Is Fizzdoc free?
Yes, and it stays free: no sign-up, no watermark, no daily limit and no premium tier. Fizzdoc is open source under the Apache-2.0 license, so anyone can run their own copy.
More private tools
- Cut & Split AudioTrim an MP3 or M4A, or cut it into parts. Listen before you save.
- Merge AudioJoin MP3 or M4A files into one, in the order you choose.
- MP3 to TextTranscribe an MP3 recording into text.
- Video to TextTranscribe the speech in an MP4 or WebM video into text.
- Subtitle GeneratorCreate SRT or VTT subtitles from a video or audio file.
- Merge PDFCombine several PDFs into one, in the order you choose.
- Split PDFExtract the pages you need into a new PDF.
- Rotate PDFTurn sideways pages the right way, permanently.