ABOUT THE WORKFLOW

Edit and combine images with AI

Upload one or two images and describe exactly what you want.

Use a single image to edit clothing, backgrounds, lighting, or objects. Add a second image to combine scenes, insert a subject into a new environment, or borrow elements from another image.

Perfect for portraits, marketing creatives, social media, product photography, concept art, and commercial visuals.

Model

Mage Flow Edit Turbo (Microsoft)

A fast open-source image editing model designed for instruction-based image editing. It understands natural language prompts to modify existing images, combine multiple references, and generate realistic edits in only a few sampling steps.

Qwen3-VL

Provides visual understanding of your reference images, allowing the workflow to accurately interpret multiple inputs and follow complex editing instructions.

Flux2 Klein Consistency LoRA

Enhances editing consistency, helping produce cleaner composites and more reliable results when combining multiple reference images.

HOW IT WORKS

Step 1. Upload your main image

Choose the image you want to edit.

Works great with:

• Portraits

• Product photos

• Characters

• Landscapes

• Fashion photography

Step 2. Add a second image (optional)

Upload another image if you'd like to:

• combine two photos

• place someone into a new scene

• transfer objects

• use another background

• merge multiple references

Skip this step if you're editing only one image.

Step 3. Describe your edit

Write exactly what you want.

Examples:

Place the person from image 1 inside the castle from image 2.

Replace the background with a snowy mountain landscape.

Change the outfit to a black leather jacket.

Add the sports car from image 2 beside the person in image 1.

Step 4. Hit Run

The workflow generates a realistic edited image based on your instructions.

Preview the result and download your final image.

Perfect for:

Photoshop • Marketing • Social Media • Product Visuals • Advertising • Creative Design

First time? Leave every setting as-is. The default configuration is optimized for fast, high-quality edits.

RECOMMENDED SETTINGS

Standard Editing (Recommended)

• Resolution: 720 × 1280

• Steps: 6

• Guidance: 1

• Seed: Random

A balanced starting point for most edits.

Need more detail?

Increase Steps to 8–10 for more refined textures and cleaner outputs.

Generate faster

Reduce Steps to 4–5 for quick previews.

Change the output shape

Adjust Width and Height to match your project.

Examples:

• 1:1 — Social posts

• 16:9 — Banners

• 9:16 — Shorts & Reels

• 4:5 — Instagram

Repeat the same result

Use a fixed Seed instead of Random.

Prompt Tips

Clearly describe how the images should be combined or edited.

Good prompts include details about:

• environment

• clothing

• objects

• lighting

• camera angle

• composition

Example:

Place the person from image 1 on the bridge from image 2 under cinematic cloudy lighting with realistic stone textures and shallow depth of field.

Specific prompts consistently produce better results.

USE CASES

🌍 Background Replacement

Move a subject into an entirely new environment without manual editing.

🖼️ Image Compositing

Combine multiple images into one seamless composition.

👕 Outfit Changes

Replace clothing while keeping the overall scene natural.

📸 AI Photography

Create editorial, travel, lifestyle, or fashion images from existing photos.

🛍️ Product Marketing

Place products into realistic advertising scenes and promotional visuals.

🎨 Creative Image Editing

Modify objects, environments, lighting, or compositions using simple text instructions.

WHAT WORKS BEST / WHAT TO AVOID

✅ Works Great

• Clear, high-quality images

• One or two reference images

• Detailed prompts

• Background replacements

• Scene compositing

• Product edits

⚠️ May Produce Softer Results

• Low-resolution inputs

• Heavy image compression

• Very crowded scenes

• Contradictory prompts

• Tiny faces or distant subjects