A Python script to automatically rename Mac screenshots using OCR and local LLM summarization.

• Scans your Desktop for new screenshots (PNG files starting with "Screen").
• Extracts text from screenshots using Tesseract OCR.
• Summarizes the content using a local LLM (Ollama + Llama 3).
• Renames screenshots to a concise summary, preserving the original timestamp.
• Logs both old and new filenames for tracking.
• Designed to run automatically (e.g., via cron).

• Python 3.8+
• Tesseract OCR
• Ollama with the llama3 model
• Python packages: pillow, pytesseract, requests

git clone https://github.com/yourusername/rename-screenshots.git
cd rename-screenshots

Create a new virtual environment (replace venv with your preferred name):

python3 -m venv venv

Activate the virtual environment:

• On macOS/Linux:

  source venv/bin/activate

• On Windows:

  venv\Scripts\activate

With your virtual environment activated, run:

pip install -r requirements.txt

On macOS (using Homebrew):

brew install tesseract

On Ubuntu/Debian:

sudo apt-get install tesseract-ocr

• Download and install from Ollama.
• Pull the Llama 3 model:

  ollama pull llama3

• Start Ollama (if not already running):

  ollama serve

By default, the script processes PNG files on your Desktop. If you want to change the directory, edit the SCREENSHOT_DIR variable in screenshot_renamer.py.

Activate your virtual environment if not already active:

source venv/bin/activate

Then run:

python screenshot_renamer.py

1. Find the full path to your Python executable in the virtual environment:

   which python
   # Example output: /Users/yourusername/rename-screenshots/venv/bin/python

2. Find the full path to your script:

   pwd
   # Example output: /Users/yourusername/rename-screenshots

3. Edit your crontab:

   crontab -e

4. Add the following line to run the script every hour (replace paths as needed):

   0 * * * * /Users/yourusername/rename-screenshots/venv/bin/python /Users/yourusername/rename-screenshots/screenshot_renamer.py
   

5. Save and exit the editor. The script will now run every hour.

The script creates a processed_files.txt file in the same directory, logging both the old and new filenames for each processed screenshot.

MIT License