0x4F
10110
<>
//
0x4F
&&
=>
10110
{}
<>
//
01
0xA3
{}
10110
<>
0x4F
&&
FF
//
All case studies
fabric.jsPipeline imageIA

Photobooth backgrounds generated in seconds

An application that automatically produces the personalised backgrounds for a photobooth, combining canvas composition, an image processing pipeline and generative AI.

ClientLe Ptit Festif

Visit the site
Screenshot of the Le Ptit Festif website
The client

An events business where every booking is unique

Le Ptit Festif works events where the photobooth is dressed in the colours of the evening: weddings, birthdays, company parties. Every booking therefore needs its own artwork.

We did not work on their public website. Our scope is the internal production tool that creates that artwork.

The challenge

Producing bespoke artwork at volume

Designing a photobooth background is a designer’s job: framing, visual hierarchy, legibility around the area reserved for the photo. Multiplied by the number of events, that work becomes the limiting factor of the business.

Automating it without collapsing into an interchangeable template means controlling composition precisely, not merely generating an image.

The solution

Controlled composition rather than raw generation

We built a chain where AI produces the visual material while composition stays governed by explicit rules. Photo areas, margins and legibility are never left to the model’s discretion.

01

Programmatic composition with fabric.js

The background is assembled on a canvas where every element — artwork, text, reserved photo area — is positioned by rules. Layout stays controlled whatever the generated content looks like.

02

Image processing pipeline

Resizing, cropping, format handling and final export are automated, producing files the photobooth hardware can use directly.

03

AI-generated visuals

AI produces the graphic elements from the event’s theme, providing original material without falling back on a limited stock library.

04

Batch production

The tool generates all the variants a booking needs in a single operation, rather than one file at a time.

Tech stack

The technologies we chose

fabric.js for canvas composition, a Node.js image processing pipeline for exports, and generative models for the visual material.

fabric.js

Node.js

Sharp

IA générative

Canvas 2D

The outcome

Bespoke artwork without the design hours

Creating personalised artwork no longer takes a design session per event. Personalisation stays a selling point without becoming a production constraint.

MPWB

Working on something similar?

Give us thirty minutes. We will tell you plainly what is feasible, at what budget and on what timeline.

Response within 24h 100% confidential No commitment