100% Free · No Uploads · No Accounts
Go on, poke it
Speech in, word-level timed cues out. A five-minute clip lands in about forty seconds, no upload bar, no queue position.
Fast is light and quick: a small one-time download that runs well on any laptop and is happiest with clear speech. Accurate is the heavyweight: a bigger download that needs a browser able to use your graphics card, and in return it handles 23 languages, thick accents and music under the voice. You choose before each run, and both stay in your browser after the first download.
Seven looks out of the box, from a plain caption bar to word-by-word highlights that move with the speech, plus your choice of font, colour, size and position on top.
Split, merge, drag and nudge cues to a tenth of a second. Snap to word boundaries, zoom, undo as far back as you like. Nothing is destructive.
A subtitle file in SRT, VTT or plain text for any editor or player, or the finished video with the captions burned in at the original resolution. All of it is made on your own computer. No watermark, ever.
STEP 1
Video or audio, read straight off your disk into browser memory. There is no upload step to sit through.
STEP 2
Pick Fast or Accurate, press Start, and watch the cues appear live while the video stays playable.
STEP 3
Pick a look, tweak any line you like, then take the subtitle file or the burned-in video.
VIDEO IN
AUDIO IN
SUBTITLES OUT
Why is it free
ShuShie has no upload, no transcription bill and no seat to sell you. The model downloads to your browser once and runs on your own computer.
No accounts. No watermark. No trial clock. And nothing leaves your computer.
Nope. There is nowhere to send it. Your footage, outtakes and all, never leaves your disk.
Your browser downloads an open speech-recognition model, OpenAI's Whisper, and runs it on your own computer: on your graphics card through WebGPU where the browser allows it, otherwise on the processor. Your file is read straight from disk into memory, and nothing is sent anywhere.
No. Nothing leaves your device, and nothing is fetched into it either, so there is no link box. Download the video, or open it from the Drive or Dropbox folder on your computer if you use their desktop apps, then drop the file.
Current Chrome, Edge, Firefox and Safari, on desktop and phone: transcribe, edit, and export subtitle files in all of them. The burned-in video export needs desktop Chrome or Edge.
Fast is about 200 MB (80 MB without a GPU), Accurate about 1.6 GB. One time each. After that it works with the wifi off.
Yes. Fast handles clear speech well; Accurate goes further with accents, cross-talk and music under the voice, in 23 languages. Unusual names and slang are worth a quick skim, and every line is editable.
Yes. Every cue drags to a tenth of a second, so you can land a punchline exactly where you want it, with unlimited undo if you change your mind.
A sixty-minute podcast is fine, it just takes a coffee. Audio runs in chunks and progress survives a tab reload.
There is not one. Nothing to sign into, nothing to cancel, no onboarding email sequence.
Most of them are watching you on mute. Say something they can read.
Drop your file anywhere
or