Medical fund fraud, where individuals fake treatment documents to solicit donations, is a growing concern in crowdfunding. Traditional verification methods are often manual, slow, and prone to error. This project introduces an AI-based system using YOLOv8 to detect text in medical bills and Paddle OCR to extract key information. Extracted data—like hospital names and treatment costs—is verified using fuzzy matching against a trusted hospital database. This automated approach enhances accuracy, blocks fraudulent requests, and helps restore donor trust.