FACEHUB AI
ServicesUse CasesPricing
Sign In
Sketch

Draw Where Things Go

Drop your product photos onto a canvas, block out the rest with rough shapes, and generate a finished image that keeps the arrangement exactly where you put it.
E-commerce Sellers
Art Directors
Marketers
Packaging Designers
Small Brands
Try It Free View Pricing
A sketch canvas: a cut-out photo of a white SUV dropped onto a white background, with green mountains and a blue road roughed in by hand around it

Your product, dropped in. Everything else roughed out.

The finished image: the same white SUV on a mountain road at golden hour, positioned exactly where it sat on the sketch

Rendered, with the arrangement kept.

The Problem

Composition is the hardest thing to say in words. 'Bottle on the left, slightly turned, with the box behind it and to the right' comes back with the bottle centred and the box gone. Add 'on a marble surface near a window' and the arrangement changes again. Every regeneration is a fresh roll of the dice on a layout you already know exactly — and for a real product, the label has to be the real label.

The Solution

Sketch is a canvas. Drop in the actual product photographs, move them where they belong, scale them, and rough in the rest with simple shapes to mark what sits where. Your arrangement is handed to the image model as a layout to honour rather than a sentence to interpret, so the finished shot has your composition and your products in it.

How It Works

1
Drop in your products

Upload the real photographs — the actual bottle, the actual box, the actual garment. Position and scale them on the canvas so they sit where the finished shot should have them.

2
Block out everything else

Rough shapes mark what the rest of the frame holds: a surface here, a prop there, a window for the light. It does not need to be neat. It needs to say where things are.

3
Say what it should look like, then generate

The canvas carries the layout, so your words carry the finish — the material, the light, the mood. Generate variations and the arrangement stays put while the styling changes.

What You Get

Your real products, not a model's impression of them

The composition you drew, kept through generation

Try many looks against one fixed arrangement

No prompt engineering for spatial relationships

Rough shapes are enough — nothing has to be drawn well

Consistent layout across a whole campaign

Works for products, interiors and full scenes

No design software required

Example Results

The sketch canvas with a product photo and hand-drawn mountains and road

Rough is fine. Position is what carries.

The rendered scene at golden hour, the car where the sketch put it

The car sits where you put it.

A generated lifestyle photograph of a woman beside a sports car

Product, place and person in one composition.

Ready to Try It?

Start with 50 free credits. No credit card required.

Get Started Free

Frequently Asked Questions

No. Rough rectangles and circles are enough — they exist to say where things sit, not to look like anything. The parts that need to be accurate are the product photos, and those you upload.

Your uploaded photograph is used as a reference for that object, so labels, shapes and colours come from the real thing rather than being invented.

Yes, and that is the main reason to work this way. The arrangement is fixed on the canvas, so you can generate the same composition in daylight, in studio light, on marble or on wood without re-describing where anything goes.

Closely, but it is guidance rather than a hard mask. Large, clearly separated elements hold their positions best; several small objects crowded together give the model more room to reinterpret.

A prompt describes a layout in a sentence and the model reconstructs one. Here the layout already exists as an image, so there is nothing to reconstruct — the model dresses an arrangement instead of guessing at one.