M.A.S.H (Multi Agent System for Hospitals) is a decentralized healthcare orchestration platform comprising a Desktop Interface for doctors/pharmacists, and a Mobile Interface for patients. The system is designed with a minimal, jargon-free UI to streamline clinical workflows and remove patient administrative friction. Upon system boot, M.A.S.H spools up 6 specialized AI agents. Using the Band of Agents SDK (BandSDK), these agents are distributed across multiple virtual, secure rooms (Reception & Location Room, Clinical Consultation Room, and Pharmacy & Inventory Room) to coordinate patient visits while protecting event traffic. 1. For Patients: Patients can book appointments from anywhere via the Mobile Interface. During booking, they interact with a chatbot regarding their symptoms. The Registration Agent automatically assigns the most suitable doctor by matching the patient’s issue with doctors’ specialties and availability. On arrival, the Patient Navigation Agent guides the patient directly to the doctor’s room. 2. For Doctors: In the Consultation Room, the Patient Summary Agent aggregates medical histories, lab results, and surgeries from Supabase. Written prescriptions trigger the Medicine/Prescription Management Agent. 3. For Pharmacy & Stock: In the Pharmacy Room, the Medicine/Prescription Management Agent audits stock level. If available, it routes the order to the pharmacy. If out-of-stock, it requests human-in- the-loop validation for alternate drug comments. Concurrently, the Stock Management Agent tracks usage trends to automate reorders. This multi-room agent mesh completely automates administrative handoffs in real-time.
Category tags:"M.A.S.H is one of the most comprehensive and ambitious healthcare submissions in the hackathon. The 3-room Band architecture (Reception/Location, Clinical Consultation, Pharmacy/Inventory) mirrors actual hospital departmental structure and effectively separates concerns while protecting event traffic. The 6-agent coverage of patient booking, doctor matching by specialty, navigation, medical history aggregation, prescription management, and automated inventory reorders addresses the complete clinical-administrative workflow. Dual interfaces (Desktop for clinical staff, Mobile for patients) show product thinking beyond proof-of-concept. Integration with Supabase for patient data, human-in-the-loop for out-of-stock drug substitutions, and automated reorder triggers for usage trends demonstrate production-grade design. 4-person team, 69 upvotes — reflects the strong community recognition for this comprehensive healthcare platform."
Sanem Avcil