Generating your image...
Generate or upload an image to get started
Generating your image...
Generate or upload an image to get started
This will remove all images, history, and recent generations. This action cannot be undone.
To edit an image, you need to select one first. Here are your options:
Once an image is displayed, click the "Select" button in the top-right corner of the canvas to select it for editing.
Core Principle
Describe the scene, don't just list keywords. The model excels at understanding natural language descriptions.
Template:
A photorealistic [shot type] of [subject], [action/expression], set in [environment].
The scene is illuminated by [lighting], creating a [mood] atmosphere.
Captured with [camera/lens details], emphasizing [key details].
Example:
"A photorealistic close-up portrait of an elderly Japanese ceramicist, deeply focused while shaping clay on a pottery wheel. Soft natural light filters through a window, creating gentle shadows. Shot with 85mm lens, shallow depth of field, emphasizing the weathered hands and concentrated expression."
Template:
A [style] illustration of [subject], featuring [characteristics] and [color palette].
The design should have [line style] and [shading style]. [Background description].
Styles to Try:
Adding/Removing Elements:
Using the provided image, please [add/remove] [element].
Ensure the change integrates naturally with existing lighting and style.
Style Transfer:
Transform this image into the style of [artist/style],
preserving the original composition but rendering with [stylistic elements].
Selective Editing:
Change only the [specific element] to [new description].
Keep everything else exactly the same.
Instead of "armor", describe: "ornate elven plate armor, etched with silver patterns, high collar shaped like falcon wings"
Use terms like: wide-angle shot, macro lens, bird's eye view, low-angle perspective, 85mm portrait lens
Start with a base image, then refine with follow-up edits: "Make the lighting warmer", "Add more detail to the background"
Gemini excels at text rendering. Be explicit: 'Display the text "HELLO" in bold sans-serif font'