Built by AJ Ahmed | Powered by Nano Banana Pro (Gemini 3 Pro Image) | Integrated with Claude Code | Open Source

Claude Code Now Generates Images

A custom skill that gives Claude Code the ability to generate and edit images directly from the terminal using Google's Nano Banana Pro model.

What this demonstrates: By creating a Claude Code skill that integrates Nano Banana Pro (Gemini 3 Pro Image), Claude Code can now generate images, edit photos, render text, and output in any aspect ratio or resolution - all from natural language in the terminal. Every image on this page was generated during a single conversation.

Generate in Any Aspect Ratio

Same prompt, different dimensions. Choose the right format for your use case.

# Add --aspect to specify dimensions
python3 generate_image.py "your prompt" output.png --aspect 16:9
1:1

1:1 Square

Icons, avatars, social posts

16:9

16:9 Widescreen

Banners, YouTube, desktop

9:16

9:16 Portrait

Stories, TikTok, mobile

4:3

4:3 Standard

Presentations, tablets

21:9

21:9 Ultrawide

Cinematic, hero banners

Generate in Any Style

Describe the aesthetic you want. Same subject rendered five different ways.

# Include style in your prompt
"mountain landscape, watercolor painting style, soft brushstrokes"
Flat

Flat Design

Vector art, simple shapes

3D

3D Render

Volumetric, glossy

Photo

Photorealistic

DSLR quality, sharp

Watercolor

Watercolor

Soft brushstrokes, artistic

Minimal

Minimalist

Clean lines, negative space

Choose Your Resolution

Balance quality and speed. 1K for thumbnails, 4K for prints.

# Add --size for resolution
python3 generate_image.py "prompt" output.png --size 2K
1K

1K (~1024px)

Fast generation, web thumbnails

2K

2K (~2048px)

Balanced, standard web use

4K

4K (~4096px)

Maximum detail, print-ready

Edit Images with Natural Language

Make precise changes by describing what you want. No masks, no complex tools.

# Use --input to edit an existing image
python3 generate_image.py "turn on the fireplace" output.png --input original.png
Original Image
Original

"a modern living room with a grey sofa, wooden coffee table, large window showing a garden, fireplace turned off"

Edited Versions (same dimensions preserved)
Remove plant
Remove Object

Plant Removed

"remove the plant from the coffee table"

Fireplace on
Change State

Fireplace Lit

"turn on the fireplace with warm flames"

Blue sofa
Change Color

Navy Blue Sofa

"change the sofa to navy blue"

Winter
Change Background

Winter Scene

"change the garden view to a snowy winter scene"

Evening
Change Lighting

Cozy Evening

"change to evening with warm indoor lighting and lamps turned on"

Render Text in Images

Include text in your prompt for logos, posters, and signage.

# Put text in quotes within your prompt
"logo with text 'QUANTUM' in bold futuristic font"
Logo

Tech Logo

"modern tech logo with text 'QUANTUM' in bold futuristic font, blue to purple gradient"

Poster

Quote Poster

"motivational poster with 'The future belongs to those who believe...'"

Menu

Menu Board

"coffee shop menu with 'COFFEE HOUSE', chalkboard style"

Quick Reference

All available options at a glance.

Option Values Example
Aspect Ratio 1:1, 16:9, 9:16, 4:3, 3:4, 21:9, 2:3, 3:2, 4:5, 5:4 --aspect 16:9
Resolution 1K, 2K, 4K --size 2K
Edit Image Path to source image --input photo.png
Style Reference Up to 14 reference images --ref style.png