Multilingual PDF OCR and Translation API

The business problem and approach

Built a versatile API tool for processing and translating PDFs across multiple languages. Leveraged the Python fitz library for content extraction and preprocessing, combined with Gemini 2.5 and Document AI for OCR capabilities, enabling accurate text recognition even from scanned documents. Integrated Google Translate and Azure Translation APIs to offer high-quality translations into the user's target language. The API supports end-to-end multilingual document workflows, facilitating seamless cross-border collaboration and global knowledge sharing.

Implementation highlights

Extracted and preprocessed PDF content using fitz (PyMuPDF), Utilized advanced OCR models (Gemini 2.5, Document AI) for robust text recognition, Delivered high-quality translations via Google Translate and Azure Translation APIs

Delivery and outcomes

Python (fitz library), Gemini 2.5 Model, OCR

Technologies used

• Improved cross-border team collaboration with real-time accurate translations

← Back to selected work