ARCHIVES

Access to understandable legal information and relevant welfare programmes can be difficult when information is distributed across legal documents, government portals and service providers. This paper presents a prototype of a Generative AI-driven legal assistant intended to help women and other applicants explore legal information, understand uploaded documents and discover potentially relevant government schemes through a single web interface. The design combines a frontend, a Flask/FastAPI-style backend, a database, a retrieval-augmented generation (RAG) component, a document-processing module and a scheme-recommendation module. User queries are validated and classified; relevant material is retrieved from a curated legal and government knowledge base; and the language model prepares a plain-language response with source-oriented guidance. Scheme recommendations are matched against profile details and eligibility conditions, while uploaded documents can be summarized to support comprehension. The supplied project artefacts include architecture, use-case, sequence, activity and data-flow diagrams, together with interface screenshots for scheme discovery, application tracking and emergency-support information. These artefacts demonstrate the prototype workflow and interface coverage, but do not establish measured legal accuracy, recommendation precision, response latency or user impact. The paper therefore reports a design-and-prototype contribution and defines a validation plan rather than claiming unmeasured performance. Safety measures include official-source prioritization, visible limitations, privacy-aware document handling and referral to qualified legal or emergency services where appropriate