Automatic FIR Generation System Using NLP and Machine Learning

ABSTRACTII. This paper presents the design and development ofan AI-based FIR Assistant that automates theidentification of relevant Indian Penal Code (IPC)sections from written complaint descriptions. Thesystem integrates Natural Language Processing(NLP) techniques with a MERN stack(MongoDB, Express, React, Node.js) webinterface to simplify FIR documentation for policepersonnel.III. The backend uses Python-based NLP modulessuch as tokenization, Named Entity Recognition(NER), and crime-intent classification to extractmeaningful information from complaint text. Theextracted features are matched with an IPCknowledge base using hybrid logic consisting ofrule-based mapping and machine-learning–supported similarity scoring. The MERN-basedfrontend enables officers to input complaints,view recommended IPC sections, and generate astructured FIR automatically.IV. This work demonstrates that NLP-assisted legalclassification significantly reduces manual errors,improves documentation speed, and enhancesaccuracy in FIR preparation, enabling moreefficient policing and digital governance.