Every frame here was taken on a real device with a real project open. No mock-ups, no composites and no finished edit posed for the camera — the films are the work being done, at the speed it happens.
Voice separation with an AI model that runs on the device, lifting the original dialogue out and leaving the music and effects. Subtitles written from what is said, in nine languages, timed to where each line is spoken. Animated caption and title styles — typing with a cursor, karaoke, a word that lights up as it is spoken, pop and fade — burned into the exported file exactly as the player draws them, or written out as .srt. Line detection that marks where every line starts and ends, the step Pro Tools calls Strip Silence, with arrows that walk you from mark to mark. Dub takes recorded against the picture and placed by a measured audio round trip, inside the EBU R37 lip-sync window. A voice strip heard live: high-pass, compressor, de-esser, three-band EQ and room reverb. Level, pan, fades and automatic ducking under dialogue. Fifteen picture looks picked from squares of your own frame — nine free, six with the subscription, four of them painted over the picture: a 1970s CRT television, film grain, a light leak and worn VHS tape, plus duotone and split tone. Each look on its own track, so one film can carry several. ITU-R BS.1770 loudness with EBU R128 true-peak limiting at −1 dBTP, delivery in stereo or discrete 5.1, and variable frame rates normalised to constant. It runs with no network at all, needs no account, and nothing is uploaded or charged by the minute. On Android phones and tablets, on iPhone and iPad from the App Store, and on macOS.
Filmed on a MacBook, a Galaxy Tab A8 and a Huawei handset, all running DubWeaver 2.21.1. The interface is in English here; the closing frames of each device show the same project in Russian, Japanese and German.