the-rabak/gemini-imagegen

Goal

Generate and edit images with the Gemini API using the correct model, settings, and file handling.

Use this skill when

• Generating images from text prompts.
• Editing or refining existing images.
• Applying style transfer, creating logos with text, building stickers, or producing product mockups.
• Running multi-turn image refinement or combining reference images.

Do not use this skill when

• You are about to use a model other than gemini-3-pro-image-preview without an explicit request.
• You are about to save Gemini's default JPEG output with a .png extension.
• GEMINI_API_KEY is not available.

Operating rules

• Require GEMINI_API_KEY in the environment.
• Default to gemini-3-pro-image-preview.
• Default to 1K resolution and 1:1 aspect ratio unless the task needs something else.
• Save default Gemini output as .jpg.
• Use conversational refinement for iterative edits.
• Use multiple reference images when composition requires them.

Procedure / Reference

Defaults

• Model: gemini-3-pro-image-preview
• Resolution: 1K by default; available values are 1K, 2K, 4K
• Aspect ratio: 1:1 by default
• Supported aspect ratios: 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9

Basic generation

import os
from google import genai
from google.genai import types

client = genai.Client(api_key=os.environ["GEMINI_API_KEY"])

response = client.models.generate_content(
    model="gemini-3-pro-image-preview",
    contents=["Your prompt here"],
    config=types.GenerateContentConfig(
        response_modalities=['TEXT', 'IMAGE'],
    ),
)

Custom resolution and aspect ratio

config=types.GenerateContentConfig(
    response_modalities=['TEXT', 'IMAGE'],
    image_config=types.ImageConfig(
        aspect_ratio="16:9",
        image_size="2K"
    ),
)

Editing an existing image

from PIL import Image

img = Image.open("input.png")
response = client.models.generate_content(
    model="gemini-3-pro-image-preview",
    contents=["Add a sunset to this scene", img],
)

Multi-turn refinement

chat = client.chats.create(
    model="gemini-3-pro-image-preview",
    config=types.GenerateContentConfig(response_modalities=['TEXT', 'IMAGE'])
)

response = chat.send_message("Create a logo for 'Acme Corp'")
response = chat.send_message("Make the text bolder and add a blue gradient")

Save the result correctly

image.save("output.jpg")

Do not do this unless you explicitly convert formats:

image.save("output.png")

Convert to PNG if required

from PIL import Image

for part in response.parts:
    if part.inline_data:
        img = part.as_image()
        img.save("output.png", format="PNG")

Verify the real file format

file image.png

If file reports JPEG image data, rename it to .jpg or resave it properly.

Prompting guidance

• Photorealistic scenes: include lens, lighting, angle, and mood.
• Stylized art: name the visual style explicitly.
• Text in images: specify font style and placement.
• Product mockups: describe the lighting setup and surface.

Advanced features

Google Search grounding:

response = client.models.generate_content(
    model="gemini-3-pro-image-preview",
    contents=["Visualize today's weather in Tokyo as an infographic"],
    config=types.GenerateContentConfig(
        response_modalities=['TEXT', 'IMAGE'],
        tools=[{"google_search": {}}]
    )
)

Multiple reference images, up to 14:

response = client.models.generate_content(
    model="gemini-3-pro-image-preview",
    contents=[
        "Create a group photo of these people in an office",
        Image.open("person1.png"),
        Image.open("person2.png"),
        Image.open("person3.png"),
    ],
    config=types.GenerateContentConfig(
        response_modalities=['TEXT', 'IMAGE'],
    ),
)