AI Processing
MixUp’s AI Wizard applies AI processing to a guest’s photo or video. You set it up once as an AI template, pin one AI mode and its parameters, and the booth renders every guest’s content through it.
The AI menu tabs
Section titled “The AI menu tabs”The AI menu has three tabs:
- AI Photo — AI-process still photos. See AI Photo templates.
- AI Video (“Ai Live”) — animate a frame into a short video. See Video AI modes.
- AI Video editing — edit the AI video in the full video editor. See AI Video editing.
The photo modes
Section titled “The photo modes”| Mode | What it does | Needs |
|---|---|---|
| Ai Painter | Repaints the photo from a prompt and settings | a prompt |
| Deep Fake | Swaps the guest’s face onto a reference image | a reference image |
| Ai Chromakey | Removes the background and drops in your background image | a background image |
| ChatGPT Image / Flux Kontext / Gemini | Prompt-driven edits by external foundation models | a prompt (paid) |
See Photo AI modes for the detail on each.
The video modes (“Ai Live”)
Section titled “The video modes (“Ai Live”)”Kling and Pixverse model families animate a frame into a clip, with length, quality, effect, and style options. See Video AI modes.
Cloud vs local
Section titled “Cloud vs local”- Cloud is the default for every AI mode — the source is uploaded to the AI service, queued, processed, and downloaded back.
- Local processing runs only for Deep Fake photo, and only if you opt in and download the models. It runs on the booth PC and costs no coins. See Local AI.
The processing sequence
Section titled “The processing sequence”Processing always runs in the same direct order:
- AI Photo only. The booth takes the photo → it’s sent to processing (Local render or Cloud — see Render) → the processed photo then appears on the Sharing screen.
- AI Photo + AI Video. The photo is taken; if AI Photo is on, the photo is processed first, and that AI photo is then passed on to AI Video for processing. If there’s no AI Photo, the original photo is sent to AI Video instead.
What one session produces
Section titled “What one session produces”With both photo and video enabled, a single session yields three final pieces of content:
- the original photo (in its frame),
- the AI photo (in its frame),
- the AI video.
You can hide the original photo with a booth setting if you’d rather not show it — see Booth Options → Sharing screen.
Demo vs paid
Section titled “Demo vs paid”Some modes run in Demo for evaluation (Ai Painter, Deep Fake, Ai Chromakey). The external models (ChatGPT/Flux/Gemini) and all video are paid — they need coins and won’t run on a demo account. See Credits & errors.
Get started
Section titled “Get started”New to AI? Follow Create an AI project for the simple step-by-step, then dive into the per-mode pages.