Skip to content

AI photo editor and image to image

Upload a photo and say what to change. The model edits only what you ask and keeps the rest of the image. Free to try: new accounts get 100 credits with no card.

Model

New and popular

All models

Google image model, edits with 14 references

Images to edit0/4 images

Settings for this model are syncing. Check back soon.

Credits come back automatically if a job fails.

Example

Image to image example
Try a prompt like Replace the background with a sunlit kitchen, keep the product unchanged

Write an edit that preserves what matters

Treat your uploaded image as the starting point. Separate the requested change from the details that should remain, and inspect the output against the original.

Product restyling

Change the surroundings or lighting of an existing product. Check label text, proportions and distinctive edges before using the result commercially.

Illustration variations

Explore a new palette or rendering style while preserving the pose and composition. Large style changes may alter details that need a second pass.

Image to image prompt starters

Suggested briefs to adapt to your project. These are starting points, not promises of a specific result. Using a prompt fills the editor without submitting a generation.

Replace a background

Replace the background with a matte warm-gray studio wall and soft floor shadow. Keep the product shape, camera angle, label and colors unchanged. Do not add objects or text.

Upload the product image first. The prompt defines both the edit and the details to inspect afterward.

Change a palette

Restyle this illustration with a muted blue and terracotta palette. Keep the character pose, facial expression, outlines and composition. Leave the background simple.

A constrained color change is easier to review than changing the pose, style and setting together.

Choose settings for the result you need

Available controls depend on the selected model. Check the displayed quote after changing settings; a starting price may describe a different configuration.

Input image
Choose a clear source at a useful resolution. Small text and obscured details are harder to preserve.
Output shape
Match the source shape when preserving framing matters; a different ratio may extend or crop the composition.
Multiple references
Only add extra images when the selected model supports them. Describe each reference’s purpose rather than expecting conflicting examples to combine automatically.

When the result needs another pass

The subject changed

Reduce the scope of the edit and explicitly name the features to preserve. Regenerate from the original rather than repeatedly editing an already-drifted result.

A logo or label changed

Compare at full size. Generative editing is not a guarantee of pixel-identical preservation. Use manual retouching where exact identity is essential.

The requested edit is ignored

Remove unrelated directions and specify the region or object in ordinary visual terms.

How it works

  1. 01

    Upload a photo

    JPG, PNG or WebP up to 10 MB. Add more images to combine them.

  2. 02

    Say what to change

    Name the part to edit and what it should become.

  3. 03

    Generate and download

    Compare the result with the original and run another edit if needed.

Image to image pricing

You pay per result in credits, priced by model and settings. The exact cost shows next to the generate button, and failed jobs are refunded automatically.

See credit packs and plans
Cheapest model
—
Free on sign up
100 credits

Image to image questions

Will the rest of the image stay the same?

Nano Banana Pro and GPT Image 2.5 keep untouched areas closely. Say clearly what must not change.

Can I combine several images?

Yes. Upload more than one image and describe how to combine them, for example a product into a scene.

Can I edit photos of real people?

Only with their consent, and never to create misleading or sexual content.

Is the image to image generator free?

Yes, you can start for free. New accounts get 100 credits with no card, and you get 10 more each day you sign in while your balance is under 50. Larger jobs use paid credits.

Do I need to sign up?

You need a free account to generate, and signing in with Google takes one click. You can browse examples and see the credit price without an account.

How much does the image to image generator cost?

Each generation shows its credit price before you start. One credit is about $0.01, and packs start at $9.90 for 990 credits.

How long does a generation take?

Most images finish in under a minute.

What happens if a generation fails?

You get the credits back. If a job fails, times out or is blocked by moderation, the credits return to your balance automatically.

Can I use the results commercially?

Yes. Files you generate with purchased credits can be used commercially, within each model provider's terms and our content policy.

Are my prompts and results private?

Yes. Prompts and results are private to your account. Results are kept for 30 days. On tools that accept uploads, source files are kept for 7 days.

What files can I upload?

JPG, PNG or WebP images up to 10 MB. Tools that take reference clips or audio also accept MP4 or MOV video up to 50 MB and MP3 or WAV audio up to 20 MB.

Can I run several generations at once?

Accounts that have purchased credits or a subscription can run up to 4 generations at a time; free-only accounts can run one. Each job shows its own progress, and finished results are saved to My creations.

Do I need an API key or another subscription?

No. You use every model through FrameMux with one credit balance, with no API key and no separate model subscription.