Audio to Text
Transcribe speech from any audio or video file with Whisper running on your device — timestamps included, export as TXT, SRT or VTT, nothing leaves your browser.
How to Use Audio to Text in 3 Steps
Drop the recording
Any audio or video your browser can play — voice notes, meetings, podcasts, lectures. The speech track is decoded on your device.
Pick a model and language
Tiny is quick and fine for clear speech; Base and Small are more accurate and handle accents and 99 languages. Leave the language on auto-detect unless you know it.
Transcribe and export
Click Transcribe. The first run downloads the model once; text streams in as it is recognised. Copy it, or download TXT, SRT, VTT or JSON.
Audio to Text — Frequently Asked Questions
Is my recording uploaded anywhere?
No. OpenAI’s Whisper model runs inside your browser with Transformers.js (WebGPU or WebAssembly). The audio never leaves your device; only the model file is downloaded — from Hugging Face, once — and cached by the browser.
How accurate is it?
Tiny is roughly what a phone’s offline dictation gives you; Base is noticeably better with accents and background noise; Small approaches cloud services on clear recordings. Names, jargon and heavy crosstalk are where all three slip — proof-read before publishing.
How long does it take?
Roughly real time to 3× faster than real time with Tiny on a laptop, slower with Small or on phones. A 30-minute podcast with Base takes about 10–20 minutes on a modern laptop with WebGPU. Keep the tab open.
Which languages work?
All 99 Whisper languages, including Hindi, Tamil, Telugu, Bengali, Marathi, Gujarati and Urdu. Translate to English gives an English transcript of speech in any of them.
Can I get subtitles for a video?
Yes — drop the video file itself, then download as SRT or VTT. Burn them in with the Video Subtitle Burner, or upload the SRT next to the video on YouTube.
Why is the first run slow to start?
The model (41–249 MB) is downloaded and compiled the first time. After that it loads from the browser cache in a few seconds. If the download stalls, check an ad-blocker is not blocking huggingface.co.