AI Photo templates

The AI menu has three tabs at the top: AI Photo, AI Video (Video AI modes), and AI Video editing. This page covers AI Photo — the templates that AI-process the photos you shoot.

You can create and select several AI templates for a project; with more than one selected, the guest gets to choose which one to use.

Create and open a template

Press New to create a template, then double-click it (or press Edit) to open the template editor.

The editor header has the usual actions: rename, copy, delete, export, tags, and open folder. The header also shows the price of this processing, and a Test button (below).

AI Mode

AI Mode is the base processing applied to the photo. The modes are:

AI Painter
Deep Fake
AI Chromakey
ChatGPT Image
Flux Kontext
Gemini (several versions; also known as Nano Banana)

AI Painter

The most flexible mode. Its extra settings (Painter only):

Compliance to the original — the balance the AI leans on: original priority, balance, or prompt priority.
Draw mode — photorealism, drawing, or caricature.
Additional options — the AI engine version, a seed, and an upscaler on/off.
Nudity filter — turn it on and if the AI produces something indecent (nudity), it won’t pass the filter and won’t be shown — it’s sent back for another attempt.
Additional styles — preset styles on the AI server; you can stack several so they interact.

Deep Fake

Swaps the face from the original photo into a reference image. Upload the Deep Fake reference — the image the guest’s face is placed into.

Deep Fake can also run locally: on a good PC it processes quickly on-device, with no internet, and it’s free (built-in). See Local AI.

AI Chromakey

Removes the background and drops in your uploaded background. It’s cheap. On a modern PC you can use the built-in AI chromakey — free, real-time, and no internet; on a weaker PC, use the cloud chromakey instead.

ChatGPT Image, Flux Kontext, Gemini

More advanced processing — you can give more complex prompts, and they draw on current information for more up-to-date results. Great for tailored work, for example brand customization with survey tags.

Deep Fake to fix faces — a sub-option for these: some AI slightly loses a person’s features, so this extra pass finds the face in the final image and corrects it to look more like the original.
Gemini 3 Pro / 3.0 Flash — Use Google Search on/off: lets the model pull in more current, real-time data. If your prompt mentions something the model doesn’t already know, search finds it.

Image sets (image packs)

In the additional options you can enable image sets. Image sets are uploaded on your KSuse account (in Sets of Pictures — see KSuse account & portal); when you create a new set it appears here in the list, and you add it to the template.

Then reference the set explicitly in the prompt — for example “take the first photo as the original, and use the next photos as accessories,” or “use these as style references.” So a single processing uses not just one photo but the original plus extra images the AI can also draw on.

This lets you combine styles and elements — for example a “doll-packaging” look, or adding team badges, logos, and shapes. A nice extra layer of customization.

Template preview and customization files

For each mode you can upload a template preview — the image shown to the guest in the booth UI, so they see roughly what the processing looks like.

Deep Fake and AI Chromakey also take a single customization file — the reference (Deep Fake) or the background (AI Chromakey):

If you upload only the customization file (no template preview), that file is shown in the booth.
If you upload both, the template preview is shown.

The AI prompt

The most important part. Press Edit for two large fields:

Prompt — the task you set for the AI.
Negative prompt — what must not appear in the result.

Describe both in detail for a more predictable result. You can insert substitution tags from a survey to customize the result from the guest’s answers.

A bar at the top shows the tags available from other areas — other AI prompts, the templates for this event, and the surveys for this event — so you don’t have to look them up.

Test

From the editor you can Test the AI processing: pick a photo and it’s sent to the AI with your current settings.