AI image generation

Create images from a prompt, a reference, or both.

ImageStyle is an AI image generator for moving from an idea to a usable visual. Start with a written direction, add a reference image when the selected model supports it, then choose the framing and output settings that fit the job.

Open image generatorExplore focused tools
Two ImageStyle example outputs: a beverage campaign poster and a sofa product-detail image set.Two ImageStyle example outputs: a beverage campaign poster and a sofa product-detail image set.
Examples from the current ImageStyle product and campaign workflows.
Start with
A prompt or supported reference image
Create
New image variations in a selected frame
Control
Model, aspect ratio, and output settings

Examples to inspect before you generate

Inspect different image directions before choosing a model, frame, and prompt structure.

Every card identifies whether it is a verified model-family example or a visual reference. Reference cards are not model benchmarks or before-and-after claims.

Editorial poster showing clothing draped over a blue chair on a pale background.

Poster composition

One central object, clear negative space, and a deliberate text hierarchy.

Verified model-family example

Verified GPT Image family example from the local ImageStyle prompt library.

An isometric miniature landmark scene on a floating base with a simplified landscape.

Illustrated place concept

Combine a place, its defining landmarks, and a clear viewing angle in one visual direction.

Verified model-family example

Verified Nano Banana family example from the local ImageStyle prompt library.

A sculptural food-based typographic composition on a wooden serving board.

Concept-led object treatment

Give an abstract idea a material, a physical form, and a simple composition rule.

Verified model-family example

Verified GPT Image family example from the local ImageStyle prompt library.

What the image generator is built to help with

Text-to-image directions

Describe the subject, setting, composition, and visual mood you need, then turn that direction into a fresh image.

Reference-guided variations

When the selected model accepts reference images, use one to give the generation a visual starting point alongside your prompt.

Purposeful framing

Choose from the aspect-ratio and output options available for the model instead of designing around a single fixed canvas.

A practical AI image workflow

The workspace keeps the creative choices together so the prompt, reference material, model, and output settings can be adjusted in one place.

  1. 1

    Describe the result you need

    Write the visual direction in your own words, including the subject, composition, and the details that matter most.

  2. 2

    Add a reference when it helps

    Use a supported reference image when you need the generation to begin from a particular visual direction.

  3. 3

    Select available output options

    Choose a model, aspect ratio, and output setting that are available for the result you want to make.

  4. 4

    Generate, inspect, and refine

    Review the visible result, then adjust the direction or settings before creating another variation.

What to check before you publish a result

Model capabilities vary
Reference-image support, available aspect ratios, output tiers, and credit cost depend on the model and settings you select.
Review visible details
Generated images should be checked for product details, readable text, hands, brand marks, and any other element that must be accurate.
Use publishable source material
Only upload images and references that you are allowed to use in your intended project.

AI image generator FAQ

What can I create with the AI image generator?
You can create a new visual direction from text, or make a reference-guided variation when the selected model accepts reference images. The result is shaped by your prompt, the model, and the available output settings.
Can I use the AI image generator for free?
New accounts receive welcome credits for eligible 1K image generation. Higher-resolution image generation and video generation require purchased credits, and model costs vary.
Can I use a reference image?
Yes, on models that support reference images. The workspace shows the upload option and the model-specific limits when that capability is available.
Do all image models offer the same settings?
No. Supported aspect ratios, output tiers, reference-image limits, and credit cost vary by model. Choose the model first, then use the settings it makes available.
Can I use the output for product or campaign work?
You can create visuals for those workflows, but review the finished image before publishing. For product-specific work, use a supplied reference image and verify every detail that must remain accurate.

Updated:

Capability source: Current ImageStyle image-generation workspace.

Open image generator