gemini-cli

name	gemini-cli
description	Generate images and text using Google Gemini API via CLI - image, text, models. Use when user mentions 'generate image', 'create image', 'Gemini image', 'Nano Banana', 'Imagen', 'AI image generation', 'image editing', 'describe image', 'text generation with Gemini', or wants to create/edit/analyze images using the Google Gemini API. Supports Nano Banana, Nano Banana Pro, Nano Banana 2, and Imagen 4 models.
category	devtools

gemini-cli

When To Use This Skill

Use the gemini-cli skill when you need to:

• Generate images from text prompts (Gemini native or Imagen models)
• Edit existing images with AI (change colors, remove backgrounds, add elements)
• Describe or analyze images using Gemini vision
• Generate text with Gemini models (with optional file inputs)
• List available Gemini/Imagen models and their capabilities

Setup

If gemini-cli is not found, install and build it:

bun --version || curl -fsSL https://bun.sh/install | bash
npx api2cli bundle gemini
npx api2cli link gemini

api2cli link adds ~/.local/bin to PATH automatically.

Always use --json flag when calling commands programmatically.

Working Rules

• Always use --json for agent-driven calls so downstream steps can parse the result.
• Start with --help if the exact action or flags are unclear instead of guessing.
• Image generation can take 30-120 seconds depending on model and complexity.
• Output images are saved to ./out/ by default. Use --out <path> for a specific file.

Authentication

gemini-cli auth set "your-google-api-key"
gemini-cli auth test

Auth commands: auth set <token>, auth show, auth remove, auth test

Token is stored in ~/.config/tokens/gemini-cli.txt.

Resources

image

Command	Description
gemini-cli image generate --prompt "A cat in space" --json	Generate image with default model (gemini-2.5-flash-image)
gemini-cli image generate --prompt "Logo design" --model gemini-3-pro-image-preview --aspect-ratio 1:1 --json	Generate with Nano Banana Pro
gemini-cli image generate --prompt "Landscape" --model gemini-3.1-flash-image-preview --image-size 2K --json	Generate with Nano Banana 2, 2K size
gemini-cli image generate --prompt "Photo" --model imagen-4.0-generate-001 --sample-count 4 --json	Generate 4 images with Imagen 4
gemini-cli image generate --prompt "Product shot" --images-only --out ./product.png --json	Images only, specific output path
gemini-cli image generate --prompt "Art" --aspect-ratio 16:9 --image-size 4K --json	Custom aspect ratio and size
gemini-cli image edit --prompt "Make the sky purple" --input ./photo.jpg --json	Edit an existing image
gemini-cli image edit --prompt "Remove background" --input ./product.png --out ./clean.png --json	Edit with specific output
gemini-cli image describe --input ./photo.jpg --json	Describe/analyze an image
gemini-cli image describe --input ./diagram.png --prompt "What does this show?" --json	Analyze with custom prompt
gemini-cli image models --json	List available image generation models

Image generation models:

• gemini-2.5-flash-image - Nano Banana (default, fast)
• gemini-3-pro-image-preview - Nano Banana Pro (high quality)
• gemini-3.1-flash-image-preview - Nano Banana 2 (newest)
• imagen-4.0-generate-001 - Imagen 4
• imagen-4.0-ultra-generate-001 - Imagen 4 Ultra (highest quality)
• imagen-4.0-fast-generate-001 - Imagen 4 Fast

Aspect ratios: 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9

Image sizes: 512, 1K, 2K, 4K

text

Command	Description
gemini-cli text generate --prompt "Explain X" --json	Generate text with default model
gemini-cli text generate --prompt "Summarize" --input ./doc.pdf --json	Generate with file input
gemini-cli text generate --prompt "Analyze" --input ./a.png ./b.jpg --json	Generate with multiple file inputs
gemini-cli text generate --prompt "Translate" --model gemini-2.5-pro-preview-06-05 --json	Use specific model
gemini-cli text generate --prompt "Write code" --system "You are a senior engineer" --json	With system instruction
gemini-cli text generate --prompt "Short answer" --max-tokens 100 --temperature 0.5 --json	With generation config
gemini-cli text models --json	List available text models

models

Command	Description
gemini-cli models list --json	List all available Gemini API models
gemini-cli models list --filter image --json	Filter models by name
gemini-cli models list --filter imagen --json	List Imagen models only
gemini-cli models get gemini-2.5-flash-image --json	Get details about a specific model

Output Format

--json returns a standardized envelope:

{ "ok": true, "data": { ... }, "meta": { "total": 42 } }

Image generation returns:

{ "ok": true, "data": { "model": "...", "prompt": "...", "images": ["./out/123-1.png"], "texts": ["..."] } }

On error: { "ok": false, "error": { "message": "...", "status": 401 } }

Quick Reference

gemini-cli --help                         # List all resources and global flags
gemini-cli image --help                   # List image actions
gemini-cli image generate --help          # Show image generate flags and models
gemini-cli text generate --help           # Show text generate flags
gemini-cli models list --help             # Show models list flags

Global Flags

All commands support: --json, --format <text|json|csv|yaml>, --verbose, --no-color, --no-header

Exit codes: 0 = success, 1 = API error, 2 = usage error