Unlimited transcription
A ten-second clip or a two-hour interview, as many as you like. The only limit is how fast your Mac is, and it is faster than you expect.
Buy it once and it is yours: unlimited videos, unlimited captions, and every one of them written on your own Mac.
Everything the app does, with nothing metered and nothing held back.
A ten-second clip or a two-hour interview, as many as you like. The only limit is how fast your Mac is, and it is faster than you expect.
Every word carries its own timestamp, snapped onto the real edge of the sound. Measured at about 3 ms against known speech onsets — roughly twenty times tighter than the browser version manages.
Runs Whisper Large v3 Turbo by default — 574 MB, against the 78 MB model the browser version starts with — with Metal acceleration and real voice-activity detection. Fewer wrong words means less to correct.
SubRip (.srt), WebVTT (.vtt), styled SubStation (.ass) for burn-in, and word-level JSON that keeps every timestamp. What the preview shows is what libass renders.
Give it the proper nouns and jargon before you start — a client's name, a product, a place — and the transcript uses your spelling instead of guessing at it.
English, Spanish, French, German, Italian, Portuguese, Dutch, Polish, Russian, Japanese, Korean, Chinese, Hindi, Arabic, Turkish and Ukrainian. Leave it on detect and it works out which one it heard, tells you which, and writes the subtitles in it — or translates any of them into English.
Most captioning tools ask you to upload your footage so their servers can transcribe it. suubbs does the transcribing on your machine, so the question of what we do with your video never comes up — we never receive it.
That is the complete list. There is no telemetry, no usage reporting and no analytics anywhere in the app — not switched off by default, simply not written.
On first launch it downloads the speech model — 574 MB, once, from Hugging Face — and keeps it in your Application Support folder. After that you can turn the wifi off and it will caption all day. That download is the app fetching something from the internet; nothing of yours goes the other way.
You do not have to take our word for any of this. The free version in your browser runs the same code, and a browser tab cannot quietly phone home — open the network inspector and watch it stay empty while it transcribes. Try that now →
Stripe takes the payment. You land back here with a download link, and it also goes to your email so you can find it again later.
Drag suubbs into Applications. It is signed with a Developer ID and notarised by Apple, so it opens on a double-click without any warnings to click through.
It fetches the speech model once — 574 MB, a minute or two — and then it is yours offline. Drop in a video and hit Generate.
Told straight, because finding out afterwards is how you end up wanting a refund.
Download it, drag it to Applications, open it. Everything it needs to read your video is inside the app — no Homebrew, no command line, no separate downloads except the speech model it fetches once on first launch.
Universal binary: Intel and Apple Silicon both run natively. Apple Silicon is considerably quicker, because transcription runs on the GPU through Metal.
The engine is portable and Windows is planned, but it is not built yet. If that is what you need, don't buy this today.
Export a styled .ass file and ffmpeg burns it into the picture. There is no one-click burn-in inside the app yet.
You pay once and every 1.x update is included. If there is ever a 2.0 worth charging for, this copy keeps working exactly as it does today — it does not phone home to check whether you are still allowed to use it, because it does not phone home at all.
Email hello@suubbs.com and we'll refund you. There is no licence to deactivate and no seat to reclaim.