The voice moves the mouth
The picture is generated against the real speech — not guessed first and dubbed afterwards.
A crew, a studio and a presenter who never misses a line — all on your desk.
The picture is generated against the real speech — not guessed first and dubbed afterwards.
Two or three pictures of one place become an edit. Cuts land between sentences, never mid-word.
Any script length, filmed in clips and joined seamlessly. Change one sentence — only that clip is filmed again.
Gender, age, pitch, pace, energy, warmth and tone — or clone a recording. Ten languages, one consistent voice.
Pick the displays in your scene. They come out flat chroma green, presenter in front, ready for your graphics.
The app reads the depth of every scene. Desks, chairs and audiences stay in front of the green.
News studio, keynote stage, late-night set. Tune size, screens, light and style — drawn in minutes.
Each movie shows its quality. Keep the best one — a quick draft never overwrites it.
From composed news anchor to animated keynote speaker. Add a handheld camera or a slow push-in.
A listening model compares every sentence with the script. A missing word, and it is spoken again.
Several stills per camera. The one most like your presenter is kept.
No account, no upload, no per-minute bill. Once the models are downloaded, it works in airplane mode.
Create projects and render movies from the command line — for batch jobs, your own scripts and long renders over SSH.
How script mode works →A picture of the place, a few photos of the person, the script. Press Make Movie, go to bed, and wake up to a finished movie — made on your Mac.
Free to try: film your first two shots. Unlock the whole movie with a one-time purchase. Apple silicon with 128 GB of memory (32 GB works), macOS 26 or later. The models download once; nothing is sent anywhere.