Requirements and FAQ
What your Mac needs, which models do the work, how long a movie takes — and answers to the questions about likeness, privacy and licences.
What it needs
A capable Mac and a few hundred gigabytes of disk. In return, a studio that never sends a bill.
Your Mac
- Chip
- Apple silicon
- Memory
- 128 GB recommended, 32 GB works
- System
- macOS 26 or later
- Free disk
- about 200 GB
- Languages
- 16, follows your Mac
The models
- Film
- MiniMax H3
- Placing the presenter
- FLUX.2 Klein
- Voice
- Qwen3-TTS
- Seeing
- Qwen2.5-VL · Whisper · Qwen3
How long it takes
- Voice and stills
- minutes
- Draft shot
- about 4 min
- Standard shot
- about 25 min
- High shot
- about 2½ h
- Sharper small faces
- + about 30 min a shot
Good questions
What does it cost?
Free to try: film your first two shots. Unlock the whole movie with a one-time purchase. Every project films its first two shots for free, with a small mark in the corner and a closing card. The purchase unlocks every shot of every project, without them, and is shared with your family. Bought the app before it was free? Everything stays unlocked.
Can I make a movie of anyone?
Only of yourself, or of someone who agreed to it — the same goes for cloning a voice. The app asks you to confirm this for every presenter, and every movie says in its metadata that it was generated. Don't make people say things they never said.
Does anything get uploaded?
No. The models are downloaded once from Huggingface. After that, every picture, voice and movie is made on your Mac — it works offline.
How good is the lip sync?
Good in medium and close shots at Standard quality, looser at Draft, where a face is only a few pixels tall. The picture is generated against the real voice, but this is a general video model, not a dedicated lip-sync tool.
Which languages can the presenter speak?
English, German, French, Spanish, Italian, Portuguese, Russian, Japanese, Korean and Chinese — in the same voice.
Which languages does the app itself speak?
English, German, French, Spanish, Italian, Polish, Ukrainian, Dutch, Portuguese, Swedish, Danish, Norwegian, Finnish, Czech, Slovak and Hungarian. The app follows your Mac's language — separately from the language your presenter speaks.
Why does it take so long?
Every frame is generated on your Mac by a film model with tens of billions of parameters. Draft is quick enough to check an edit; Standard renders about a minute of film overnight, High a few shots.
Can I use the movies commercially?
That depends on the model licences. FLUX.2 Klein 9B is for non-commercial use only — switch to Klein 4B (Apache 2.0) in the Models window for commercial work. The film model comes under the MiniMax H3 Community License; read it before you publish. The Models window lists every licence.