Abstract: To address multi-source data, complex rules and stringent intranet security constraints in hydropower test reporting, a closed-loop generation system based on Flask/Vue was designed. The system integrates template rendering, rule checking, OCR/rule-based extraction, and review traceability. During the pilot phase, the time required to prepare chemical test reports was reduced from 60 minutes to 8 minutes, and the review time was reduced from 10 minutes to 2 minutes. The time required for automated network-related testing was reduced from 300 minutes to 120 minutes. The results demonstrate that the system can improve report standardization and preparation efficiency and is suitable for private deployment on corporate intranets.