Published Paper Details:

In the linguistically diverse areas, access to basic healthcare information is scarce as it is affected by language barriers, delayed medical care, and unavailable digital support tools. This paper presents a multilingual healthcare assistant for preliminary symptom guidance, emergency severity assessment, and nearby hospital discovery across Indian languages. The system combines a web-based interface with machine learning models and rule-based reasoning to provide reliable and safety-aware answers. It has various Indian languages that allow interaction with users of different backgrounds. The architecture is a React based frontend with Fast API based backend, which guarantees scalable and real-time performance. Experimental evaluation demonstrated 90% symptom classification accuracy, 0.88 macro F1-score, and end-to-end response latency below 320 ms. It is a non-diagnostic design system and is focused on emergency escalation during the high-risk situations. Findings suggest that with the use of multilingual AI-based assistants in healthcare, preliminary healthcare support can be substantially more accessible and responsive.