Skip to main content
Create takes your reference images, uses Gemini to write an optimized prompt, then sends that prompt to FLUX.2 to generate product photos. You don’t write prompts yourself — the AI builds them from what you upload.
Create page showing upload slots for Product, Model, Outfit, Pose, and Background alongside the results grid

​
How it works

1

Upload reference images

Add images to any of the five category slots: Product, Model, Outfit, Background, and Pose. Product is the only required slot. The rest are optional.
2

AI refines the prompt

When you click Generate, GESTEL sends your images to Gemini. Gemini analyzes each one and writes a structured prompt that describes your product, the scene, and any people or styling you’ve provided.
3

FLUX.2 generates images

The refined prompt goes to FLUX.2 (Pro or Fast engine, your choice). Images appear in the results grid as each one finishes — you don’t have to wait for the full batch.
4

Download what you like

Hover any result to reveal the download button. You can also open an image in the lightbox to preview it at full size before saving.
Results grid showing some images still loading and others fully generated

​
Create vs. Recreate

Both modes generate images with the same engines. The difference is what the AI does with your inputs.
CreateRecreate
What you provideReference images (product, model, outfit, pose, background)A scene photo + a product photo
What the AI doesInvents a new scene around your referencesPlaces your product into your existing scene
Best forExploring new visual directionsFitting a product into a specific real photo
Use Create when you want the AI to build the scene from scratch. Use Recreate when you already have a backdrop you want to use.

​
Engine choice

The engine you pick affects how many images you get, how much it costs, and the style of results.
EngineCreditsImages per generation
Pro108
Fast34
Studio Upload141
Pro uses FLUX.2 Pro with JSON-structured prompts — better for final or hero shots. Fast uses FLUX.2 Klein with prose prompts — better for quick iteration. Studio Upload is a special mode that requires a Learn asset and produces a single highly consistent image.

​
Next steps

Uploading references

How to add images to each slot, including from the Learn library.

Image categories

What each category does and how the AI uses it in the prompt.

Engines

Detailed comparison of Pro, Fast, and Studio Upload.

Advanced options

Camera, lighting, shot type, and aspect ratio controls.