Generate professional AI images using Google's Nano Banana models via the Gemini API.
- API key must be set as
GEMINI_API_KEY
environment variable
- Uses curl for all API calls (no SDK required)
| Model | Identifier | Best For |
|---|
| Nano Banana PRO | gemini-3-pro-image-preview
| Professional assets, text rendering, infographics, 4K output, complex multi-turn editing |
| Nano Banana | gemini-2.5-flash-image
| Fast generation, simple edits, lower cost |
Default to PRO for quality work. Use Flash for rapid iterations or simple tasks.
BEFORE calling the API, always craft an effective prompt. Read references/prompting-guide.md
for comprehensive prompting strategies. Key principles:
- Describe scenes, don't list keywords - Write narrative descriptions, not tag soup
- Use natural language - Full sentences with proper grammar
- Be specific - Define subject, setting, lighting, mood, materials
- Provide context - The "why" helps the model make better artistic decisions
- Edit, don't re-roll - If 80% correct, ask for specific changes
For any image, specify:
- Image type: What kind of visual (photo, infographic, logo, sticker, etc.)
- Content: Specific elements, data, or information to include
- Style: Visual style, color palette, artistic approach
curl -s -X POST \
"https://generativelanguage.googleapis.com/v1beta/models/gemini-3-pro-image-preview:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"contents": [{
"parts": [{"text": "YOUR_PROMPT_HERE"}]
}],
"generationConfig": {
"responseModalities": ["TEXT", "IMAGE"],
"imageConfig": {
"aspectRatio": "16:9",
"imageSize": "2K"
}
}
}'
curl -s -X POST \
"https://generativelanguage.googleapis.com/v1beta/models/gemini-3-pro-image-preview:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"contents": [{
"parts": [
{"text": "YOUR_EDIT_INSTRUCTION"},
{"inline_data": {"mime_type": "image/png", "data": "BASE64_IMAGE_DATA"}}
]
}],
"generationConfig": {
"responseModalities": ["TEXT", "IMAGE"]
}
}'
| Parameter | Values | Notes |
|---|
aspectRatio
| 1:1
, 2:3
, 3:2
, 3:4
, 4:3
, 4:5
, 5:4
, 9:16
, 16:9
, 21:9
| Match use case |
imageSize
| 1K
, 2K
, 4K
| Use uppercase K; PRO model only for 4K |
Add
"tools": [{"google_search": {}}]
to generate images based on current information:
curl -s -X POST \
"https://generativelanguage.googleapis.com/v1beta/models/gemini-3-pro-image-preview:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"contents": [{"parts": [{"text": "Create an infographic of current tech stock prices"}]}],
"tools": [{"google_search": {}}],
"generationConfig": {
"responseModalities": ["TEXT", "IMAGE"],
"imageConfig": {"aspectRatio": "16:9"}
}
}'
Use the ICS framework and prompting guide. Examples:
Photorealistic:
A photorealistic close-up portrait of an elderly Japanese ceramicist with deep wrinkles and a warm smile, inspecting a tea bowl. Soft golden hour light from a window. 85mm lens, shallow depth of field. Serene mood.
Infographic:
Create a clean, modern infographic explaining photosynthesis as a recipe. Show "ingredients" (sunlight, water, CO2) and "finished dish" (energy). Style like a colorful kids' cookbook page.
Product Shot:
High-resolution studio photograph of a matte black ceramic coffee mug on polished concrete. Three-point softbox lighting, 45-degree angle, sharp focus on rising steam. Square format.
Use
scripts/generate-image.sh
or call API directly:
./scripts/generate-image.sh "Your prompt here" output.png --ratio 16:9 --size 2K
The API returns base64-encoded image data. Extract and decode:
# Response contains: {"candidates":[{"content":{"parts":[{"inlineData":{"mimeType":"image/png","data":"BASE64..."}}]}}]}
# Extract with jq and decode:
cat response.json | jq -r '.candidates[0].content.parts[] | select(.inlineData) | .inlineData.data' | base64 -d > image.png
- Use 16:9 or 21:9 for hero images
- Specify brand colors, modern/minimal style
- Include text requirements in prompt
- Use 1:1 aspect ratio
- Request "minimalist", "clean lines", "vector-style"
- Specify color scheme explicitly
- Describe lighting setup (softbox, natural, studio)
- Mention surface/background materials
- Include camera angle and lens type
- Define data to visualize
- Specify style (corporate, playful, technical)
- Request clear text and labeled sections
- Request "bold outlines", "kawaii", "cel-shading"
- Specify "white background" or "transparent background"
- Define color palette
- PRO supports up to 14 reference images
- Explicitly state: "Keep facial features exactly the same as Image 1"
- Describe expression/pose changes while maintaining identity
See scripts/generate-image.sh
for a ready-to-use generation script.
For advanced techniques including:
- Photorealistic scene templates
- Text rendering best practices
- Sequential art and storyboarding
- Dimensional translation (2D↔3D)
- Search grounding for real-time data
Read references/prompting-guide.md
.