Picture IPTC AI Generator

Key Features

• 🔍 AI-Powered Caption Generation
  Utilizes the Salesforce/blip-image-captioning-base transformer to intelligently analyze and generate meaningful image captions.

• 🏷️ Dynamic Keyword Extraction
  Automatically generates up to 10 relevant keywords using n-gram frequency analysis from the generated caption. Improves discoverability on digital platforms.

• 📝 Title & Description Creation
  Converts AI captions into SEO-friendly titles and concise image descriptions (up to 200 characters).

• 🖼️ IPTC Metadata Embedding
  Uses iptcinfo3 to embed metadata directly into image files, including:

  • Title (object name)

  • Description (caption/abstract)

  • Keywords

  • Copyright Notice (auto-filled with current year)

• 🗂️ Organized Output
  Saves processed images in an output directory with cleaned, title-based filenames. Optionally exports associated metadata as a .json file for integration with DAM systems.

• ⚡ Fast Mode Option
  Toggle faster caption generation using shorter token limits for performance-sensitive workflows.

📂 How It Works

1. Place your .jpg, .jpeg, or .png images in the upload folder.

2. Run the script.

3. Images are analyzed, and new JPEGs are saved in the output folder with embedded IPTC metadata.

4. Optionally, metadata JSON files are saved alongside each image.

🔧 Requirements

• Python 3.11.8

• torch, transformers, Pillow, iptcinfo3

💼 Ideal For

• Photographers

• Digital asset managers

• Media organizations

• Stock image contributors

• Social media marketers

• Anyone automating metadata generation for large image sets

• Install Python, run :
• put your pictures in /upload
• iptc.py in cmd
• find picture : in output folder