Generate
Tools
Templates
Learn
Use cases
Pricing
Sign in Start creating

Image to Image

Upload an image and reshape it with a prompt while keeping the structure.

Drop a source image

On the free plan, generations are queued behind Pro and Max members and take at least 40 seconds. Upgrade for priority and faster results.

Reading your prompt

See what the image to image can do

Drag to compare: keep the composition, transform the entire style.

Source photo example for the Image to Image Restyled example for the Image to Image Source photo Restyled
Drag to compare: keep the composition, transform the entire style.

Three simple steps

1

Upload a base image

Drop in the photo or sketch you want to transform as your starting point.

2

Write a change prompt

Describe the new look you want while keeping the original composition in mind.

3

Set strength and run

Lower strength stays close to the source, higher strength changes more, then save your pick.

Convert your result

Need the finished image in another format? Convert it in the same workspace.

Want the detail? Read How the image model works, or see the img.now image model.

What image to image does

Image to image takes a picture you already have and rebuilds it through the lens of your prompt. Instead of inventing a scene from nothing, the model studies the colors, shapes and layout of your upload, then repaints them in a new direction. This makes it perfect for restyling a photo, coloring a sketch, or turning a rough mockup into something polished. The image to image guide walks through the workflow step by step.

Understanding strength and denoise

The single most important control here is strength. Think of it as a dial between faithful and free. A low setting respects your original almost completely and only nudges the style. A high setting treats your image as loose inspiration and feels closer to a fresh text to image render. Finding the sweet spot usually takes two or three tries.

StrengthWhat happensBest for
0.2 to 0.3Subtle restyle, layout intactColor tweaks, light cleanup
0.4 to 0.6Clear new style, shapes keptRestyling photos and art
0.7 to 0.9Heavy rework, loose linkSketch to finished piece

Example edit prompts

Pair each prompt with the right strength. These recipes show how to describe a change without throwing away the parts you want to keep.

Repaint this portrait as a soft watercolor, keep the pose and face, low strength
Turn this daytime street photo into a rainy night scene with neon signs, medium strength
Convert this pencil sketch into a full color fantasy illustration, high strength
Give this living room photo a warm autumn palette and cozy lamp light, low strength

Tips for cleaner results

Start with a clear, well lit base image, since the model can only work with what you give it. Keep your prompt focused on the change you want rather than redescribing the whole scene. If you want to fully reinvent the subject instead, the main generator may serve you better, and our prompt writing guide helps you phrase edits in a way the model understands. For a single targeted change, Magic Edit follows a plain-language instruction; to commit a photo to a whole new look use Restyle, or change just an object's colour with Color Replace.

Questions about the image to image

What does the strength setting do?
Strength, sometimes called denoise, decides how much the model is allowed to repaint. Low values keep your photo mostly intact, while high values let it reinvent the scene.
When should I use this over text to image?
Use image to image when you already have a layout, sketch or photo you like and only want to change the style, lighting or details rather than start from scratch.
Will it keep faces and shapes the same?
At lower strength it holds onto the overall shapes and pose. Push strength too high and the original structure starts to drift away.
Try Image to Image