For everything spoken · interviews, Shorts, tutorials, vlogs

Auto captions, made on your machine.

Word-by-word social-style captions, as a panel right inside Premiere Pro. No upload queue, no subscription, no watermark — and free forever for videos up to 10 minutes.

Live preview in the panel
29
languages, auto-detected — including Arabic and Hebrew, laid out correctly right-to-left
$0
for videos up to 10 minutes — no watermark, no account
1:1
preview and render share the same drawing routine — pixel-identical
0
uploads — NDA footage stays on your machine

From rough cut to finished captions

The panel runs inside Premiere. You export nothing, you import nothing — captions land as a transparent clip above your edit.

Video coming · 8 s
Step 1

Transcribe

Sequence open, one click. Optionally only the range between in and out points. Whisper runs locally — up to the largest model, Large-v3, even on the free plan.

Video coming · 8 s
Step 2

Correct & style

Uncertain words are highlighted: click, type, Enter. Next to it you set style, font, colors and words per cue — live in the preview.

Video coming · 8 s
Step 3

Onto the timeline

Rendering runs in the background, Premiere stays usable. Afterwards a ProRes clip with true transparency sits on the next free video track.

The details that make the difference

Remembers everything

Sessions survive

Close the panel, close Premiere — next time, transcript, corrections and style are back. Per sequence, up to eight of them.

Your fonts

Every installed font

Searchable list, each font rendered in itself. Your clients’ brand fonts included.

Panel in 10 languages

Follows your Premiere

English, German, French, Spanish, Japanese and more — automatic, switchable.

Honest markers

Uncertain words visible

Whatever recognition wasn’t sure about is highlighted — you check precisely instead of reading everything.

Non-destructive

Your edit stays untouched

No re-render of your footage, no changes to clips — just one new track on top.

Offline

Works on a plane

After the one-time model download, nothing needs the internet.

And when there’s music underneath?

Then every standard recognition fails — ours included. That’s why music mode exists: it separates the vocals from the instrumental before transcribing. How music mode works →

Your first captions are five minutes away.