Reve Edit Fast

Edit images with text prompts while keeping the original composition intact.

What it does

Reve Edit transforms existing images using natural language. Upload a photo and describe the changes you want—the model modifies your image while preserving its spatial relationships, lighting, and overall structure.

Unlike basic filters or overlay effects, this model understands how objects relate to each other in space. When you edit something, shadows fall correctly, reflections align with light sources, and perspective stays consistent. The result feels like a natural continuation of the original image, not a patched-together composite.

What makes it good

Spatial intelligence: The model understands depth, perspective, and how objects interact in three-dimensional space. This means edits blend seamlessly with the existing scene instead of looking pasted on.

Composition preservation: When you modify parts of an image, everything else adjusts appropriately. Change the lighting on one object and the surrounding shadows update. Replace a background and the subject’s lighting adapts to match the new environment.

Prompt adherence: The model follows your instructions closely. Describe what you want changed, and it makes those specific edits without randomly altering other parts of the image.

Aesthetic quality: Outputs maintain professional-looking results with proper texture, material rendering, and visual coherence.

What you can use it for

Product photography: Update product colors, swap backgrounds, or adjust lighting without reshooting. Make variations of the same product shot for different campaigns.

Photo restoration: Remove unwanted elements, fix lighting issues, or restore damaged areas while keeping the photo’s original character.

Landscape editing: Change weather conditions, adjust time of day, or modify natural elements while maintaining realistic perspective and lighting.

Creative iteration: Quickly explore different visual directions from a single starting image. Test variations without starting from scratch each time.

Character consistency: Edit images of people or characters while maintaining their unique features and proportions across different scenes or poses.

How to use it

The model takes two inputs: your reference image and a text prompt describing the changes you want. The prompt should be specific about what to modify. Instead of “make it better,” try “change the sky to sunset colors” or “remove the person on the left side.”

The more precise your description, the more control you have over the result. You can specify colors, positions, styles, or entire scene changes. The model handles both subtle tweaks and dramatic transformations.

Tips for good results

Be specific in your prompts. “Replace the background with a modern office” works better than “change the background.”

The model excels at maintaining consistency, so it’s particularly good for edits that need to look believable—not stylized or abstract transformations.

For complex edits, you might need to run the model multiple times, making incremental changes that build on each other.

Try reve/edit on playground to see how it works with your own images.