A TypeScript CLI tool for bulk thumbnail generation using Google Gemini and OpenAI image generation APIs.

Generated with OpenAI GPT Image 1.5 | 1024x1024 | quality: high

• Dual API Support: Generate images using Gemini (Nano Banana) or OpenAI (GPT Image 1.5) including their latest models as of December 2025
• Flexible Input: Text prompts, reference images, or both
• Batch Processing: Process multiple jobs from a single JSON file
• Parallel Generation: Configurable concurrency for faster processing
• Progress Tracking: Real-time progress bar and summary

git clone https://github.com/daveonkels/nano-cli-image-gen.git
cd nano-cli-image-gen
npm install

Copy the example environment file and add your API keys:

cp .env.example .env

Edit .env:

GOOGLE_AI_API_KEY=your-gemini-api-key
OPENAI_API_KEY=your-openai-api-key

Get your API keys:

• Google AI Studio
• OpenAI Platform

npm run start -- -i prompts.json -o ./thumbnails

# Generate using Gemini Flash (default)
npm run start -- -i examples/prompts.json

# Use Gemini Pro (Nano Banana Pro) for higher quality
npm run start -- -i prompts.json -m pro

# Force all jobs to use OpenAI
npm run start -- -i prompts.json -p openai

# High concurrency
npm run start -- -i prompts.json -n 5

# Validate input file
npm run start -- -i prompts.json --dry-run

Create a JSON file with your generation jobs:

{
  "defaults": {
    "provider": "gemini",
    "geminiModel": "flash",
    "size": "1024x1024",
    "quality": "high"
  },
  "jobs": [
    {
      "prompt": "A vibrant sunset over mountains"
    },
    {
      "prompt": "High-quality product photo",
      "geminiModel": "pro"
    },
    {
      "prompt": "Style this image:",
      "image": "./reference.jpg"
    },
    {
      "prompt": "Cyberpunk city at night",
      "provider": "openai",
      "outputName": "cyberpunk-city"
    }
  ]
}

# Build TypeScript
npm run build

# Run with tsx (development)
npm run start -- -i examples/prompts.json

# Watch mode
npm run dev

MIT