A written idea.
A visual starting point.

What is text-to-image AI?

A text-to-image model creates a still image from a written prompt. Use it to explore a product scene, illustration, landscape or visual concept without uploading a source image.

Write a prompt with a clear composition

1. Subject: name the main object or scene.
2. Composition: describe framing, viewpoint and the placement of important details.
3. Light and color: specify the atmosphere and palette.
4. Style: choose a coherent direction, such as product photography or an editorial illustration.

A cobalt glass perfume bottle on dark coastal rocks at sunset. Warm golden reflections, detailed water droplets, a clean unbranded label area. Close-up product photography, portrait composition.

This is a prompt draft to adapt. It is not a claim about a tested model result.

Use the selected model’s image settings

Available controls come from the model catalog. A model may offer resolution, aspect ratio, output format or quality; controls and legal combinations vary. Changing one setting may adjust another to an available combination. The creator shows the resulting Credit cost before submission.

Do not assume a setting is available on every model. Use the choices currently shown, and review any subscription requirement before creating.

Review the image before the next step

Inspect the subject, edges, small text and composition. Refine one part of the prompt at a time. Completed images can be opened at full size or downloaded, and remain accessible in My Creations.

Work from an image you already have

Choose Image to Image to supply source images and describe changes. To add motion, download the still image and upload it in Image to Video. Review plans and Credits before submitting either task.