Care route

CareRoute — AI Agents for Faster Access to Hospital Care

CareRoute is a Jac-powered healthcare coordination application that helps patients and families find a suitable hospital with available beds. Built for the Jac AI Hackathon, it addresses a common problem: during an emergency, people often do not know which hospital has the right care unit or enough capacity, while hospitals receive little advance warning about incoming patients.

A patient describes what is happening, shares a location, and can optionally upload a medical PDF. CareRoute assesses urgency, identifies the required care unit, and recommends hospitals using bed availability, distance, and estimated travel time. The results include alternatives, reasons for each recommendation, and a map with navigation support.

CareRoute uses a team of specialized agents coordinated by an Orchestrator. The agents handle triage, document summaries, hospital matching, emergency dispatch, re-planning, and clinic reminders. A live timeline makes their actions visible, showing each step, handoff, result, and fallback.

For critical cases, the application demonstrates a coordinated response through simulated ambulance requests, police route clearance, and hospital alerts. A cancellation countdown lets users stop an accidental request. Hospital staff can accept or decline incoming alerts; if a hospital declines or becomes unavailable, CareRoute automatically searches for another suitable hospital and updates the plan. A one-tap SOS option supports users who cannot type a description.

The hospital dashboard lets staff update available beds and manage surgery waiting lists. Patients can compare planned-surgery options using estimated waiting time and distance, then join a hospital’s list. The clinic workspace saves an appointment date and time together with a patient’s or guardian’s phone number, suggests an arrival time based on hospital load, and previews simulated reminder messages.

Jac and Jaseci are central to the implementation. More than 80% of the application code is written in Jac, including the backend, frontend, graph models, and agent workflows. Jac nodes and edges represent hospital capacity, patient cases, agent events, and requests. Typed by llm() functions support structured agent decisions, while Geoapify provides location search, routing, and maps. Timeouts, circuit breakers, and deterministic fallback logic help the demo continue when external services are unavailable.

CareRoute demonstrates how AI agents can coordinate a changing healthcare journey—from understanding a request to finding care and adapting when the original plan fails.

Demo scope: Hospital data is fictional. Ambulance dispatch, police requests, and SMS delivery are simulated. CareRoute provides urgency and hospital-selection support; it does not diagnose medical conditions or replace emergency services.