Readme

Reve Remix

Combine elements from multiple images into something new. Upload between one and four reference images, describe what you want, and Reve Remix merges them into a single cohesive result.

What it does

This model takes your reference images and intelligently blends them together based on your text description. Think of it like having ingredients that you want to combine into a finished dish—each image contributes specific elements, and the model figures out how to merge them naturally.

For example, you could merge a directory of photos showing models with images of new clothing designs, then update your entire online store with fresh product shots in minutes. Or combine a landscape for the environment, a character pose from another image, and a color palette from a third to create exactly the scene you’re imagining.

When to use it

The remix model shines when you need to:

Merge product images with different contexts or models
Combine visual elements from multiple sources into one coherent image
Update e-commerce photography quickly with new inventory
Build complex scenes from separate reference images (environment, pose, lighting, style)
Create variations by mixing and matching elements across your image library

How it works

Reve understands spatial relationships, proportions, and lighting direction across your reference images. When it combines them, the result feels composed and intentional rather than like a collage. Shadows fall correctly, materials respond naturally to light, and everything aligns in perspective.

The same technology that powers Reve’s main image model—known for stunning aesthetics and accurate text rendering—handles the heavy lifting here. You get professional results without needing to master complex editing tools.

About Reve

Reve’s image models are built by a small research team focused on understanding creative intent. Their approach converts images into an internal structured format that helps the model grasp spatial relationships and visual elements more precisely than traditional methods.

The models consistently rank among the top image generators for prompt adherence, text accuracy, and visual quality. What makes Reve different is how well it handles the details that other models struggle with—readable text, multi-character scenes, and maintaining consistency when working with multiple references.