Qwen Image 2.1: Create and Edit Images with AI

Name the main person, object, or scene and describe the action or purpose.

Specify framing, camera angle, background, and where important elements should appear.

Mention materials, color palette, medium, lighting, and mood.

Quote the exact wording and describe its placement, language, and typography.

Identify the change you want and explicitly ask to preserve the subject, pose, product details, or other elements.

When providing several images, explain what each reference contributes, such as subject identity, clothing, object, or style.

Ask for a transparent background for a cutout or overlay, and verify the exported file retains an alpha channel.

Create a product concept or edit a product photo while describing which product details to preserve.

Draft a visual layout, specify headline text, and describe the intended placement and style.

Change the background, lighting, or surrounding scene while using a reference to guide identity.

Create a character or object cutout intended for a sticker, overlay, or composite.

Combine multiple subject or style references and explain the role of each image.

Describe a character, environment, art direction, and framing to explore a visual concept.

Use a mask or visual annotation with a concise instruction to focus an edit on one area.

Explore packaging directions, labels, materials, and scene styling as early-stage design concepts.

Generate an image from a written prompt describing the subject, setting, visual style, and details.

Provide an image and explain the change you want in natural language.

Combine multiple people, objects, or style references in one editing workflow.

Use circles, painted annotations, or masks to identify areas for a targeted change.

Generate or edit transparent images with an RGBA channel when the task calls for a clear background.

Reference-guided editing is designed to better preserve the identity of people and products.