AI-powered FIR generation system on GCP ingests multilingual voice and text complaints from citizens through React/React Native apps, routing them through a Speech-to-Text, NLP, and LLM processing pipeline. The AI Engine extracts incident information, detects missing details, and generates intelligent follow-up questions before drafting FIRs with legal context retrieved from a RAG system containing BNS, BNSS, and BSA documents. Police officers review, verify legal sections, and approve FIRs via a dashboard, with all data persisted in MongoDB and Cloud Storage for audit and case tracking. This architecture demonstrates how generative AI and retrieval-augmented generation solve complex domain-specific document generation while maintaining legal compliance and human oversight.