Magic Edit
Refine product shots, replace backgrounds, and apply complex visual changes while keeping subject identity and composition stable.
6 model options
Pick by style, speed, and quality target.
Prompt-to-output flow
Clear examples with practical prompts.
One subscription
All your AI apps in one platform.
What you can create
- Use one or more reference images for controlled visual edits.
- Apply local and global changes with natural language instructions.
- Switch editing models based on realism, style, and turnaround speed.
Why creators use it
- Repurpose existing assets into new campaign variants quickly.
- Cut back-and-forth in revision cycles for design teams.
- Keep editing and generation in the same production workflow.
Magic Edit examples
Reference-based editing examples that combine subjects and environments into polished, coherent final visuals.
Combine subject and vehicle into one cinematic scene
Turn two separate photos into a lifestyle shot without a location shoot or custom photography session.
Inputs
Input 1 · Subject

Input 2 · Car


Edit prompt
make a photo of the man driving the car down the california coastline
Input generation prompts
- Use provided subject photo input (already available).
- Use provided car photo input (already available).
Place a studio portrait into a real-world setting
Create natural lifestyle content for social posts using one portrait and one location reference.
Inputs
Input 1 · Person portrait

Input 2 · Coffee shop interior


Edit prompt
Create a photorealistic image of the woman sitting by the window inside this coffee shop, holding a ceramic mug, natural morning light on her face, candid lifestyle photography style, keep her facial identity and clothing consistent.
Input generation prompts
- Photorealistic waist-up portrait of a young woman in a simple neutral sweater, clean studio lighting, camera-level framing, natural skin texture, shallow depth of field, no background objects.
- Photorealistic cozy coffee shop interior with wooden tables, large window light, warm tones, empty scene, eye-level composition, high detail.
Build a dynamic product ad from multiple references
Produce high-impact ad creative by combining product detail, human motion, and environment in one generation.
Inputs
Input 1 · Sneaker product shot

Input 2 · Runner legs action shot

Input 3 · Night city street background


Edit prompt
Generate a photorealistic sports ad scene by placing the sneaker onto the runner's moving foot in this rainy neon city street, emphasize speed and motion, sharp product detail with cinematic lighting, vertical ad composition.
Input generation prompts
- Photorealistic product photo of a modern white running sneaker, side profile, floating on neutral gray background, soft studio shadows, high detail.
- Photorealistic lower-body action shot of a runner mid-stride, motion blur in legs, athletic clothing, dynamic angle, no visible face.
- Photorealistic wet city street at night with neon reflections on pavement, cinematic perspective, empty background, no people.
Transform a simple scene into an event-ready visual
Turn a plain location photo into polished marketing imagery for event invites and social announcements.
Inputs
Input 1 · Backyard daytime photo


Edit prompt
Transform this backyard into a cozy nighttime birthday party scene with warm string lights, decorated table, balloons, subtle bokeh, and a celebratory atmosphere while preserving the original layout.
Input generation prompts
- Photorealistic suburban backyard in daytime, simple patio and grass, neutral composition, no people, clear sky.
Available models
Grok Imagine Image Edit
Image EditingxAI’s prompt-based editing model for targeted image updates with support for multiple references.
xAI · January 28, 2026
Open model pageNano Banana Edit
Image EditingGoogle’s original Nano Banana editing workflow for prompt-based updates using one or more reference images.
Google DeepMind · October 2, 2025
Open model pageNano Banana 2 Edit
Image EditingFast prompt-based image editing workflow for controlled updates using one or more reference images.
Google DeepMind · November 20, 2025
Open model pageNano Banana Pro Edit
Image EditingEditing workflow based on Google DeepMind’s Nano Banana Pro image generation and editing model.
Google DeepMind · November 20, 2025
Open model pageSeedream 5 Lite Edit
Image EditingByteDance’s fast Seedream 5 Lite editing workflow for multi-reference prompt-based image updates.
ByteDance Seed · February 13, 2026
Open model pageFLUX.1 Kontext [pro]
Image EditingBlack Forest Labs’ context-aware editing model for fast iterative image transformations.
Black Forest Labs · May 29, 2025
Open model page