Automated Document OCR: Extracting ID/Passport Data into Structured Format

A complete Optical character recognition (OCR) pipeline for extracting data from official documents (national ID cards and passports) and converting it into structured JSON, ready for database storage or integration with other systems.

The system takes a document image as input, extracts text via OCR, and for passports parses the MRZ (Machine Readable Zone) for validated data. The attachment shows a real example: input document image alongside the actual structured JSON output (name, address, national ID, birth date, etc.).

Built as a FastAPI service with dedicated endpoints per document type.