Image GenerationActive
reve / remix v2
remix v2
Reve Remix V2 blends multiple reference images into one new image guided by an instruction. Use it for product mockups, scene mashups, and style-driven variants.
Image to ImageImage Editing
Model ID
remix v2
Provider
reve
Updated
1782406223
wiro playground—reve/remix v2
Updated 1782406223
## Overview
Remix V2 by Reve is an image-to-image remix model. You provide several reference images plus an instruction that explains how to merge them. The model synthesizes one new image that blends subjects, style cues, and scene elements into a single composition. It’s useful when you want a cohesive result without manual compositing.
## What you can build
- Product lifestyle shots by combining a product photo with a separate background reference
- Campaign variations from a small set of brand images and an art direction instruction
- Character-and-scene mashups that borrow pose from one image and environment from another
- Moodboard-to-frame concepts that mix palette, texture, and lighting references
- “Same idea, new setting” variants by swapping the background reference image
## Inputs
- One or more reference images you want to remix into a single output. Use common image formats like JPEG, PNG, GIF, WebP, TIFF, or AVIF.
- A written instruction describing what to keep from each reference and how to combine them. Keep it explicit about roles like “subject from image 1” and “background from image 2.”
- An optional output aspect ratio choice. You can let the model decide automatically, or pick a fixed ratio such as 16:9, 9:16, 3:2, 2:3, 4:3, 3:4, or 1:1.
## Outputs
- One remixed image file.
- The output follows the selected aspect ratio. Common standard outputs are about 1.5 megapixels: - 16:9: 1536 × 864
- 3:2: 1536 × 1024
- 4:3: 1536 × 1152
- 1:1: 1024 × 1024
- 3:4: 1152 × 1536
- 2:3: 1024 × 1536
- 9:16: 864 × 1536
## Limitations
- The model may change fine identity details like faces, logos, or small text. Don’t assume pixel-perfect preservation.
- Conflicting references can fight each other. Mixed lighting, camera angle, or perspective can reduce coherence.
- Low-quality inputs increase failure risk. Very small images, heavy JPEG artifacts, watermarks, screenshots with UI overlays, or blurry scans can cause glitches.
- The output can include invented details. Treat results as synthetic and review before publishing.
## Safety & compliance
- You must have rights to use the reference images you upload. Don’t submit copyrighted or trademarked material you don’t control.
- Don’t use it to create illegal content, exploitative sexual content, or non-consensual impersonation.
- Don’t generate hate, harassment, or content that promotes violence or self-harm.
- Don’t attempt to bypass safety checks or platform restrictions. Repeated abuse can lead to access limits or suspension.
Example prompts
Great starting points for remix v2.
Place the man and the woman to the bench in the imageImage Generation
Change the bird with the owl in the second imageImage Generation
Add yellow pickup in the second image to the grass fields in the first imageImage Generation
Add the cyclist in the second image to the path in the first imageImage Generation
API quick start
Run remix v2 with a single API call.
POST https://api.wiro.ai/v1/Run/reve/remix v2
{
"prompt": "Place the man and the woman to the bench …",
"inputImage": "https://your-cdn.com/input.png",
"aspectRatio": "auto"
}