Natural-language image editing

AI Photo Editor

Describe the change in plain language, add reference images when useful, and keep the rest of the photograph under control.

Upload a primary photo and references

First image: source · Up to 13 optional references · JPG, PNG, WEBP

Preservation

Precision edit preview

A focused object and material change while the surrounding scene remains stable.

AI photo editing before and after example

One editor for precise visual changes

Say what changes. Keep control of what does not.

A dedicated instruction field and preservation presets make the editing intent explicit before generation begins.

01

Edit with a direct instruction

Describe an object replacement, color update, material change, cleanup, or composition adjustment in everyday language.

  • Object and material edits
  • Color and background changes
  • Local cleanup and additions
Before and after precision edit of a product chair

02

Protect the regions you are not editing

Choose whether subject identity or scene structure deserves stricter protection, then reinforce it in the instruction.

  • Stable subject geometry
  • Unchanged crop and background
  • Consistent shadows and perspective
Controlled product photo edit with a stable subject

03

Use references when words are not enough

Attach supporting images for a material, object, palette, or styling direction without replacing the primary source.

  • One primary source image
  • Up to thirteen visual references
  • Fast iterations from the same base
Multiple controlled AI image editing iterations

Three-step workflow

How to edit a photo with AI

Treat the first upload as the source, then describe one clear visual change.

01

Upload the source first

Choose the photo you want to edit and add optional supporting references afterward.

02

Describe the exact change

Name what should change, what should remain fixed, and which reference to use.

03

Generate and review all regions

Inspect both the edited area and the untouched subject, scene, shadows, and crop.

Choose a more focused edit

Use a specialized workflow when the change is specifically about light, wardrobe, or camera position.

Before and after AI photo relighting example

Fix Photo Lighting

Repair flat, harsh, or mismatched light with direct controls for direction, softness, and intensity—without rebuilding the photograph.

Fix the lighting
Three outfit variations on the same fashion model

AI Clothes Changer

Upload a person first, then add garment references to build a new look without losing the model, pose, or original scene.

Change an outfit
Nine camera angles of a consistent robot character

3D Camera Control

Move the virtual camera around one image while protecting the subject's design, materials, scene, and visual continuity.

Change the camera

AI Photo Editor questions

What can I edit?

You can request object, material, color, background, lighting, styling, cleanup, and composition changes.

How should I write the instruction?

State the target change first, then name the details that must remain unchanged. Keep unrelated requests out of the same edit.

What are reference images used for?

They can communicate a product, material, palette, style, or subject that is difficult to describe precisely in words.

How many images can I upload?

The tool accepts up to fourteen images: one primary source and up to thirteen optional references.

Why did an untouched area change?

Generative edits rebuild pixels rather than locking them mathematically. Use a protection preset and explicitly name critical details.

AI photo editing before and after example

Turn a clear sentence into a controlled image edit.

Upload the source, describe one change, protect what matters, and review the result in the same workspace.

Edit a photo