Handles PDF, image, and audio using pdf2image, OpenCV, and Vosk for seamless uploading and processing

Uses Tesseract OCR and OpenCV pre-processing for accurate regional script extraction (Nepali, Sinhalese)

Leverages pydub for noise removal and Vosk for high-accuracy speech-to-text conversion

Integrates Hugging Face Transformers for contextual, fluent Nepali/Sinhalese to English translation

Employs DistilBART for concise AI-generated summaries of lengthy documents

Next.js and Material UI deliver an intuitive frontend for document upload, results display, and export

Uses PostgreSQL for scalable, privacy-focused storage of data and translation history

All core functionalities operate fully on local device, ensuring privacy, speed, and reliability without internet

Modular design with MCP tool support for easy plugin and expansion into other platforms and AI systems

Enables feedback-based fine-tuning via PyTorch for smarter, ever-evolving AI performance