image-anonymizer 0.1.4

A command-line tool to detect and mask sensitive content in images
Documentation

Image Anonymizer

A command-line tool to detect and mask sensitive content in images.

Features

  • Detects and masks sensitive content in images:
    • Text detection via OCR
  • Identifies sensitive textual information like:
    • API keys
    • Email addresses
    • Phone numbers
    • Credit card numbers
    • Personal names
    • Company or service names
  • Masks detected content with colored rectangles
  • Outputs processed images to a specified directory

Installation

cargo install image-anonymizer

Configuration

The application requires API keys to access Google Cloud Platform services. Create a .env file in the root directory based on the .env.template file:

# You need to set these environment variables
# GCP API Key for Google Cloud Platform need to access Gemini API and Cloud Vision API
GCP_API_KEY=
GEMINI_MODEL=gemini-2.0-flash-lite

Fill in your GCP API key to enable text detection capabilities.

Usage

privacy-masker [OPTIONS] <INPUT_FILE>
Options:
  -o, --output-dir <DIR>     Output directory for processed images [default: ./output]
  -m, --mask-texts <TEXTS>   Additional texts to mask, comma separated
  -h, --help                 Print help
  -V, --version              Print version

Examples

# Process a single image 
image-anonymizer screenshot.png
# Process an image and specify output directory
image-anonymizer --output-dir ./masked_images screenshot.png
# Process an image and mask additional text
image-anonymizer --mask-texts "secret,confidential" screenshot.png

Running Tests

# Run all tests excluding integration tests that require API keys
cargo test

# Run all tests including integration tests
# (requires valid API keys in .env)
cargo test -- --include-ignored

# Run specific test
cargo test test_mask_text_with_annotations

# Run tests with logging
RUST_LOG=debug cargo test

Note: Some tests require a valid GCP API key to be set in the environment or in a .env file. Tests that make API calls are marked as #[ignore] and can be run with the --include-ignored flag.

License

This project is licensed under the MIT License - see the LICENSE file for details.