How Image Generation Works

When you create a project in FrameKit, you can choose to either define a prompt and use FrameKit’s AI system to generate your image, or create an empty project and have complete control of the process yourself.

From prompt to image

Your prompt is processed in two stages:

1. Scene generation

The AI reads your prompt and identifies properties for the background and environment of your image — things like the objects, lighting, camera position, as well as more in-depth properties like surface finish (matte, glossy, metallic), transparency, and texture. These details are fed into our image generation system to output a visually striking scene with good composition.

2. Product placement and lighting

The AI then interprets the environmental details in your prompt — lighting direction and intensity, background elements, surface type, camera angle — and places/poses your product accordingly, with production ready lighting from the get-go.

What makes a good prompt

Clear, descriptive prompts produce better results. Describe:

• Lighting — direction, quality (soft/hard), colour temperature
• Background — colour, material, setting (studio, outdoor, lifestyle)
• Surface — what the product is resting on or placed near
• Mood — overall aesthetic (minimal, warm, editorial, etc.)
• Camera angle — direction/angle, position relative to the product, etc.

For more detailed guidance, see Writing Prompts - but generally speaking, the more detailed you can be, the better (especially if you provide reference images!).

Manual creation (DIY)

If you decide you want full control of your design process, our creative studio has all the tools and features you’ll need to start generating amazing images. From AI Image generation tools, to automatic lighting workflows, we have everything you need to ensure a result you can be proud to show off.