What is image to image AI?

Image to image AI begins with an uploaded reference instead of a blank prompt. In MakeImg, image to image uses that reference with your written instruction to guide a new result. You can describe a new setting, visual treatment, atmosphere, or presentation while keeping the source visible as the starting context.

What makes a useful reference image?

For image to image, choose a source that clearly shows the subject, layout, or visual quality you want to start from. A product shot, rough sketch, previous result, mood reference, or composition study can work when paired with a specific change. You should have the right to use any uploaded image.

How should I write an image to image prompt?

State the primary image to image change first, then add details that matter for the new direction. Name the setting or medium and mention lighting, palette, surface, mood, or framing when relevant. Generate a result, review it against the request, and revise the wording for the next pass.

Can I select a model and aspect ratio?

Yes. Image to image presents models and ratios compatible with the current reference workflow. Available ratios depend on the selected model, so the composer offers supported choices. Select an available ratio or use Original when it is offered to preserve the uploaded reference ratio.

What can I do after a result is generated?

After image to image completes, you can download the result, generate another direction, or open its creation workspace with the original context nearby. Use the comparison with the reference to decide what the next prompt should change, then continue image to image from the strongest version.

How are credits shown for reference edits?

Image to image displays the current credit cost in the composer before generation. Fast uses 10 credits, Standard uses 20, and Pro uses 40. Account and Checkout surfaces confirm the refresh or expiry rules for the credits you use, so you can review the cost before submitting.

Image to Image

Image to image gives an existing visual a deliberate next direction. Upload a reference, describe what should change, choose from the compatible model and ratio options, then review the generated result beside the intent you started with. This image to image workflow keeps the source, instruction, and generation controls together instead of asking you to rebuild a visual idea from a blank prompt.

0 / 10000

Model guide

Eight featured model guides, one focused workspace.

Explore eight featured model guides—GPT Image 2, Nano Banana 2, Nano Banana Pro, Seedream 5 Lite, Ideogram V3, Imagen 4 Fast, Flux 2 Pro, and GPT Image 1.5—from the same studio. Nano Banana 2 Lite and the legacy Nano Banana are also available in the generator without featured guides. Choose the model whose output best supports your brief, then continue with the result instead of rebuilding the project.

Direct each change from a visible starting point

Use image to image when the subject, composition, material, or visual language already matters. Each image to image decision stays connected to the reference and the result you want to inspect next.

Anchor image to image editing

An image to image request begins with the visual context you already have: a product, sketch, layout study, illustration, or completed result. Add that image before writing the instruction so image to image has a clear starting point. A legible reference makes it easier to judge whether the subject, arrangement, and useful details remain present while the requested direction changes. Use only images you have permission to upload, and choose a source that makes the intended focus easy to see.

Direct image to image prompts

Write an image to image prompt as a specific change, not a loose collection of style words. Name what should become different, then describe the details that support that choice: a quieter background, warmer editorial light, a matte surface, or a new illustration medium. Image to image works best when the prompt explains both the desired treatment and the visual context worth retaining. Review the first result, then adjust one meaningful instruction before asking image to image for another variation.

Choose image to image models

Image to image model choices can vary once a reference is attached. Read the model information in the composer, keep the prompt and reference visible, and choose the option that fits the edit you are making. The image to image controls present choices supported by the active workflow rather than asking you to guess whether a setting applies. If you change a model, reread the compatible options before generating so the request remains grounded in what that image to image workflow can use.

Set image to image format

The image to image ratio menu responds to the current model and reference workflow. Select an available square, landscape, portrait, or Original option according to where the result will be used. Original preserves the uploaded reference image ratio when it is offered. Keeping the ratio with the image to image request makes the composition decision visible before generation and gives you a more useful basis for comparing a restyle, scene edit, or alternate presentation.

A focused image to image workflow

Move through four visible decisions, then continue from the result that best serves the next image to image revision.

Step 1: Upload for image to image

Open the composer and add the image that establishes the image to image starting point. It may be a product photo, an interior, a rough concept, an illustration, or a prior result. Once the reference is present, image to image uses a compatible reference workflow instead of a text-only request. Check that the subject and composition are understandable before continuing, because a clear source gives you a better way to assess the generated change.

Step 2: Plan image to image changes

Tell image to image what should change first: restyle the scene, revise the environment, develop a sketch into a product concept, or shift the mood. Then add only the details that affect the outcome, such as indirect window light, a paper texture, restrained styling, or catalog framing. A focused image to image instruction is easier to compare and refine than a long list of unrelated terms. Keep the change concrete enough that you can recognize it in the result.

Step 3: Review image to image settings

Before submitting image to image, review the available model and ratio choices for the current reference workflow. Select an option that matches the destination, or use Original when it is available and preserving the reference ratio is important. Image to image shows the generation cost in the composer before you generate: Fast uses 10 credits, Standard uses 20, and Pro uses 40. This keeps the request, its supported settings, and its visible cost in one place.

Step 4: Continue image to image

Generate the image to image request, then compare the result with the source against the change you asked for. Look at the environment, material, styling, mood, or visual medium rather than expecting every detail to remain fixed. Download a completed image, generate another direction, or reopen its creation workspace to continue image to image with the original context nearby. Small prompt revisions turn each next pass into an intentional comparison instead of a restart.

Four ways to develop a visual direction

Images illustrate visual directions for image to image. They do not verify before-and-after results or state their generation source.

Image to image product restyle

Use image to image with a product image or early mockup when the object should remain recognizable while its presentation changes. Ask for a calmer palette, a different material setting, or an editorial scene. Image to image keeps the discussion anchored in the object you have while the prompt defines the campaign direction you want to explore.

Refine an existing creative direction

When a concept is close but incomplete, use image to image to describe the next adjustment instead of reconstructing the whole scene. Request a more restrained background, a seasonal atmosphere, a new illustration treatment, or a more focused composition. Image to image makes each revision easier to review against the original intent.

Develop a sketch into a visual brief

A sketch, wireframe, or mood reference can give image to image a useful arrangement and emphasis before a polished asset exists. Pair it with a prompt that identifies the medium, setting, and audience. Image to image then gives a team a visible reference and written instruction to discuss, rather than an unexplained generated visual.

Create channel-specific variations

Start image to image from one approved direction and explore related versions for a landing page, product detail, social post, or presentation. Choose a supported ratio appropriate to the destination and state the requested change plainly. Image to image helps you create considered alternatives that still relate to the source idea without forcing every output to match.

Image to Image FAQ

Bring your next reference into the composer

Use image to image to turn a clear source and a focused instruction into the next visual direction. Start where the original idea is already visible, then make each change deliberate.

Start Creating Now View Pricing Plans