Fuser

Qwen-Image Edit

Provided by Fal — Learn More

Qwen-Image Edit is a specialized image editing tool with multiple operations including: removing white backgrounds and adding realistic scenes, generating full portraits from face crops, creating retro-style group photos, seamlessly integrating products into backgrounds, generating cinematic scene transitions, removing unwanted elements, normalizing lighting, and applying designs to shirts. Each operation uses specialized LoRAs to maintain high quality and natural integration. Perfect for e-commerce, product photography, creative editing, and cinematic workflows.

Preview

Inputs

Prompt
text

The prompt describing the desired edit or effect. For next-scene operations, start with "Next Scene:" and describe camera/scene changes. For other operations, describe the background, clothing, setting, or elements to add/modify.

optional: true
Images
image
Accepts multiple

Images to edit or create variations from. Single image for most operations. Use 2 images for shirt-design (person + design). Maximum of 4 images for group-photo to combine subjects.

Operation
dropdown

The specialized editing operation to perform. Each uses a different LoRA optimized for specific tasks: background removal, portrait generation, element removal, product integration, scene transitions, lighting adjustment, retro effects, or apparel design.

default: add-backgroundAccepts: Add Background, Face to Full Portrait, Group Photo, Integrate Product, Next Scene, Remove Element, Remove Lighting, Shirt Design
Negative Prompt
text

The negative prompt is used to guide the model to avoid generating images that contain certain elements or concepts.

optional: true
Image Size
dropdown

The size of the generated image. If set to "Auto", the size of the input image will be used.

default: autoAccepts: Auto, Square, Portrait 4:3, Portrait 16:9, Landscape 4:3, Landscape 16:9
Steps
slider

The number of inference steps to perform.

default: 6•minimum: 2•maximum: 50
Seed
seed

The same seed and prompt will output the same image every time.

default: 3974
CFG Scale
slider

Controls how closely the generated image adheres to the text prompt you provide. Accepts values between 0 and 20.

default: 1•minimum: 0•maximum: 20
Effect Strength
slider

The strength of the LoRA effect to apply. 0 for no effect, 1 for full effect. For next-scene operations, recommended range is 0.7-0.8. Higher values create stronger transformations, lower values preserve more of the original image.

default: 1•minimum: 0•maximum: 1
Block NSFW
toggle

Prevents the model from generating images that portray nudity, violence, and other unsafe content.

default: false
Output Format
dropdown

The file format for the output image.

default: jpegAccepts: PNG, JPEG, WebP

Outputs

Generated Image
image

The generated image from the model.