Subtitles that fit the moment
Convert caption formats, shift timings, and inspect lines that move too quickly.
Work with subtitles ↗Sound, in writing.
Transcribe audio without uploading it. Your recording becomes editable text and subtitles, right here in your browser.
I mentioned to the universe, sir, I exist.
00:00:04.650
Recording inputLocal Whisper studio
Drop your audio or video hereOr choose a recording to get started.
Tiny transcription · 40.8 MB CPU / 119.7 MB GPU weights. Runtime files download too.
Choose a file or record your microphone.
Download and processing settings
Model weights: 40.8 MB for WebAssembly; 119.7 MB for GPU. Configuration, tokenizer and runtime files add a smaller download. Source: Hugging Face file sizes, checked 2026-10-30.
Model file sizes and source · Last checked October 30, 2026
Tiny uses the least memory. Base and Small may improve recognition but need more memory and time. Try a short sample first. No accuracy or speed is guaranteed.
Audio is processed locally. Model files download from Hugging Face when you start, then stay in your browser cache when storage permits. How privacy works
Your transcript
Click a word or timestamp to listen. Search updates as you type. Edit text and add speaker names yourself. Speaker recognition is not automatic.
Shape your subtitles
SRT and VTT exports use this layout. These controls limit characters and lines, not reading speed. Check subtitle reading speed before publishing.
Saved on this device
Recordings and edits are stored in this browser when local storage is available. Export an important transcript before clearing browser data. The delete control clears app storage and its model cache. Use browser cache controls to remove any browser-managed download copies too.
No audio uploadsNo accountNo minutes cap
A recording becomes a useful draft.
Nothing to send away. Nothing to sign into. You stay in control of the words.
Listen
Bring your recording
Choose audio or video, or record a microphone. Your browser decodes the sound locally.
Review
Give the draft a human pass
Whisper produces text and timestamps. Listen back, correct names, and add speaker labels yourself.
Keep
Take your words with you
Download a document or subtitle file. Your recording and edits can stay in this browser until you delete them.
From a first draft to a finished file.
Practical tools for the details that make a transcript usable.
Know your language’s limits
See the original Whisper paper’s language results before choosing a model.
Explore language benchmarks ↗Plan the recording
Calculate speaking time, storage size, and transcription cost using visible inputs.
Open the calculators ↗Local processing
The model comes to your audio.
MurmurScribe downloads Whisper model files from Hugging Face and runs them on your device. Your recording and transcript are not sent to a transcription server. Model downloads still require a connection and reveal ordinary connection information to the download provider.
There is no product minutes cap. Your device’s memory, browser storage, and processing speed still matter. Start with Tiny and a short recording if you are unsure. A larger model can take more time and may not fit on a phone.
Automatic text is a draft. Background noise, accents, overlapping voices, and names can cause errors. Always listen back before quoting, publishing, or using a transcript for an important decision.
Read the full privacy detailsInside the example
The speech is LibriSpeech sample 1272-141231-0000. The picture shows frequency measurements from that audio. The subtitles and word timings are an uncorrected draft exported from this app using Whisper Tiny with WebAssembly. Pressing play replays that captured result; it does not run a new transcription. Tiny heard “I mentioned” where the source says “A man said.” This is why the editor keeps the recording beside the draft.
Audio: LibriSpeech by Vassil Panayotov, Guoguo Chen, Daniel Povey and Sanjeev Khudanpur, from LibriVox reader 1272. CC BY 4.0 source · Speech excerpt. The excerpt is unchanged; the spectrogram and transcription were generated here. Last checked October 30, 2026.