A simple tool to edit or create images using various models via API (OpenAI, Google, Black Forest Labs, Bytedance, Sourceful, Stable Diffusion XL, etc.) via a terminal user interface (TUI).

• Direct image path input via CLI.
• Interactive file selection if no path is provided.
• Batch processing: Process multiple images with the same model and prompt.
• Dynamic cost calculation and display.
• Selection of Resolution and Quality.
• Pre-made and custom prompts.

1. Install dependencies:

   pip install -r requirements.txt

2. Create a .env file with your API Keys:

   OPENAI_API_KEY=your_openai_key
   GOOGLE_API_KEY=your_google_key
   OPENROUTER_API_KEY=your_openrouter_key
   OVH_AI_ENDPOINTS_ACCESS_TOKEN=your_ovh_token

Run the script:

python imgedit.py [options] [image_paths...]

# Single image (interactive selection if no path provided)
python imgedit.py
python imgedit.py photo.jpg

# Text-to-Image mode (no input image)
python imgedit.py --free

Pass multiple images to process them all with the same model, resolution, quality, and prompt:

# Multiple specific images
python imgedit.py photo1.jpg photo2.jpg photo3.jpg

# Glob pattern (all JPG files)
python imgedit.py "*.jpg"

# Glob pattern (all PNG files in a folder)
python imgedit.py ./folder/*.png

# All images in a directory
python imgedit.py ./my_folder

# Combine multiple inputs
python imgedit.py img1.jpg "*.png" ./other_folder

When you provide 2+ images, the script automatically enters batch mode:

• Shows total cost estimate (per image × number of images)
• Processes each image sequentially
• Reports success/failure count at the end

• PCX, PNG, JPG/JPEG, BMP, TIFF, WEBP

• OpenAI: gpt-image-1.5, gpt-image-1-mini. Official documentation lists prices that don't match reality (like 2-3x more expensive than what's supposed to be), so I'm doing trial&error to find the actual prices.
• Google (Nano Banana) (direct or via OpenRouter):
  • gemini-2.5-flash-image: $0.04 per image.
  • gemini-3.1-flash-image-preview: $0.07 (1K), $0.10 (2K), $0.15 (4K).
  • gemini-3-pro-image-preview: $0.15 (1K/2K) or $0.25 (4K).
• OpenRouter:
  • bytedance-seed/seedream-4.5: $0.04 per image (any size).
  • black-forest-labs/flux.2-klein-4b: $0.014 (1K), $0.017 (2K).
  • black-forest-labs/flux.2-flex: Output $0.06 (1K), $0.24 (2K); Input $0.06/MP.
  • black-forest-labs/flux.2-pro: Output $0.03 (1K), $0.075 (2K); Input $0.015/MP.
  • black-forest-labs/flux.2-max: Output $0.07 (1K), $0.16 (2K); Input $0.03/MP.
  • sourceful/riverflow-v2-fast: $0.02 (1K), $0.04 (2K).
  • sourceful/riverflow-v2-pro: $0.15 (1K/2K), $0.33 (4K).
• OVH AI Endpoints:
  • stabilityai/stable-diffusion-xl-base-1.0: Free (Rate limited: 2 per minute without API key, 400 per minute with API key). Fixed 1024x1024. With such generous rate limits it does not need API keys, but if needed you can read how to get one

Note: Black Forest Labs models cap output at 4MP, so requesting higher resolutions is useless. Input images >4MP are automatically downscaled.

This is licensed under GPL-3.0-or-later