Transform your document workflow without compromising security. SnapDoc AI processes everything on-device, ensuring your sensitive information never leaves your control.
- Your Data Stays Yours: In a world where data security is paramount, SnapDoc AI never sends your documents to external servers
- Work Uninterrupted: Process documents even without internet connection - perfect for secure environments and travel
- Enterprise Confidence: Meet compliance requirements with auditable, on-device processing
- Natural Interaction: Control the application with voice commands and listen to documents - all processed locally
- Accessibility Built-in: Voice commands and text-to-speech make documents accessible to everyone in your organization
- Intelligent Document Analysis: Instantly extract key information and insights
- Powerful Summarization: Save hours by generating concise summaries of lengthy documents
- Smart Information Extraction: Automatically identify dates, names, amounts, and contractual terms
- Voice Command Control: Navigate and analyze documents hands-free with natural voice instructions
- Text-to-Speech Conversion: Listen to your documents while multitasking
- Multi-Format Support: Seamlessly handle PDF, Word, plain text, and images (with on-device OCR)
- Enterprise-Ready Architecture: Built with structured logging, flexible configuration, and comprehensive error handling
SnapDoc AI leverages cutting-edge AI technology that runs entirely on your device:
- Qualcomm AI Hub models optimized for Snapdragon devices
- On-device machine learning for document understanding
- Privacy-preserving voice recognition
- ONNX Runtime with DirectML acceleration
- Modern software architecture following enterprise best practices:
- Structured logging with contextual information
- Flexible configuration management
- Comprehensive error handling
- Extensive unit testing coverage
This application demonstrates the power of on-device AI using:
- Qualcomm AI Hub models
- Snapdragon X series processors
- Windows on Arm optimization
- ONNX Runtime with DirectML acceleration
- Modern Python best practices:
- Structured logging with structlog
- Configuration management with dynaconf
- Comprehensive error handling
- Unit testing with pytest
# Clone the repository
git clone https://github.com/devraftel/snapdoc-edge-ai
cd snapdoc-edge-ai
# Install with uv
uv pip install -e .
# For development (includes testing tools)
uv pip install -e ".[dev]"
# Run the application
python -m snapdoc_ai
# Create and activate a virtual environment
python -m venv venv
source venv/bin/activate # On Unix/Mac
venv\Scripts\activate # On Windows
# Install dependencies
pip install -e .
# Run the application
python -m snapdoc_ai
The project includes a Makefile with common development commands:
# Install development dependencies
make dev-install
# Run the application
make run
# Run tests
make test
# Run tests with coverage
make coverage
# Format code
make format
# Lint code
make lint
# Clean build artifacts
make clean
SnapDoc AI can be configured in several ways:
- Default settings: Located in
src/snapdoc_ai/default_settings.toml
- User settings: Create a
settings.toml
file in~/.snapdoc-ai/
- Environment variables: Set variables with
SNAPDOC_
prefix (e.g.,SNAPDOC_DEBUG=true
)
- Windows on Snapdragon device (or any modern system for development)
- Python 3.11+
- For OCR: Tesseract OCR installed (for image-based documents)
MIT License