Start with a clear source image
Use a sharp photo or drawing where the main subject is easy to see. The model can only keep what it can make out: a blurry or very dark picture leaves it guessing.
Upload a picture, describe what should change, and get a new version in another style, with a new background or with edited details. The tool opens on Image to Image, ready for your upload.
Image-to-image works best when the model knows what to keep and what to change. Your picture sets the composition and the subject; your text tells the model what to do with them.
Use a sharp photo or drawing where the main subject is easy to see. The model can only keep what it can make out: a blurry or very dark picture leaves it guessing.
Name both in the prompt: "keep the person and the pose, replace the background with a city street at night". Without the first half, the model may redraw more than you wanted.
If the result drifts too far from your picture, ask for fewer changes in one request, add "keep the composition", and generate again.
Upload a photo, a sketch or a screenshot and use it as the starting point for a new image.
Turn a photo into a watercolor, an anime frame, an oil painting or a 3D render. Name the style in the prompt; the subject and layout stay close to your original.
Move the subject of your picture into a new setting: a beach, a studio backdrop, a snowy forest. Describe the new scene and the lighting you want.
Change a color, add or remove an object, or show the same scene in another season or time of day, without describing the whole picture from scratch.
The tool runs in your browser and uses Google's Nano Banana 2 by default, with other image models one click away.
Nano Banana 2 is selected by default. You can also pick Nano Banana Pro, FLUX.1 Kontext Pro or Seedream 4.5 from the model list and compare their results on the same picture.
The tool opens on the Image to Image tab with the upload ready. Switch to the Text to Image tab when you want to create something from a prompt alone.
Everything happens in the browser on a computer or a phone. When the image is ready, download it.
Three steps: upload a picture, describe the change, then generate and download.
Both live in the same tool. Which one to start with depends on whether you already have a picture to work from.
You have a product photo, a sketch or a snapshot, and you want a new version of it. The model takes your composition and subject as the base, so the result stays recognizable and you only describe what should change.
There is no source picture, or you want something completely new. Describe the subject, setting, style and lighting on the Text to Image tab, and the model builds the image from scratch. You can then upload that result on Image to Image and refine it further.
A few common ways to use image-to-image, with prompts you can adapt.
Upload a photo of a mug on a white background and ask for "on a wooden table by the window, soft morning light" to get a shot for a product card or an ad.
Upload a pencil sketch and ask for "a colored digital illustration". The model finishes the drawing while keeping your lines and layout.
Upload a photo from a trip and ask for "watercolor style, soft colors" to turn it into a picture for a card or a print.
Upload a summer photo of a house and ask for "the same scene in winter, snow, warm light in the windows" for a holiday card or a seasonal post.
One account gives you several image models, plus chat, video and music, in the browser.
Try the same picture and prompt with Nano Banana 2, Nano Banana Pro, FLUX.1 Kontext Pro or Seedream 4.5 without moving between sites.
On paid plans one monthly pool of credits covers chat, image, video and music generation, so you do not need a separate subscription for each.
Sign-up is free and needs no credit card. New accounts get welcome credits for images, video and music.
There is no app to install: YesChat runs in the browser on computers and phones, including Android and iPhone.