A "Smart Convalescence Agent" for post-surgery patients.

Inspiration

Recovery doesn't end at the hospital exit; for many, the most dangerous phase is the first 30 days at home. We noticed a "care gap" where patients feel overwhelmed by medication schedules and vital tracking, often missing early warning signs of complications. We wanted to build something that isn't just a digital diary, but a proactive partner that "thinks" like a nurse.

What it does

AuraCare is an agentic healthcare ecosystem. Using a Flutter mobile interface and IoT sensors, it monitors vitals (Heart Rate, SpO2, Temperature) in real-time.

The "Agentic" Difference: If a patient's vitals show a trend toward infection (e.g., rising temp + increased heart rate), the AI Agent doesn't just send a notification. It autonomously cross-references the patient’s recovery plan, prepares a summary for their doctor, and asks the patient specific triage questions to validate the risk.

How we built it

Mobile Frontend: Built with Flutter for a seamless, high-performance UI that handles real-time data streams from Bluetooth-connected IoT devices.

Backend & API: Developed using Laravel, providing a secure, EMR-compliant (Electronic Medical Record) structure for storing patient history and medication logs.

Intelligence Layer: A Java-based processing engine (or an LLM-integrated agent) that analyzes incoming data packets to identify health anomalies using predictive algorithms.

Connectivity: Utilized MQTT protocol for low-latency communication between the IoT hardware and the cloud.

Challenges we ran into

Data Noise: Filtering out "false positives" from IoT sensors (e.g., a spike in heart rate just because the patient walked to the kitchen). We solved this by implementing a "Context-Aware" logic that checks movement data before triggering alerts.

Latency: Ensuring the Flutter app stayed responsive while processing heavy streams of real-time health data.

Privacy: Navigating the strict data encryption requirements necessary for medical-grade software.

Accomplishments that we're proud of

Developing an Agentic Workflow where the AI can "decide" when a situation is an emergency versus a routine deviation.

Creating a highly accessible UI in Flutter that is easy for elderly post-surgery patients to navigate.

Successfully integrating a Laravel backend that can scale to handle thousands of concurrent data streams.

What we learned

We learned that in healthcare, Trust is the primary feature. It’s not enough for the AI to be right; it has to be explainable. We spent significant time ensuring the "Agent" explains why it is recommending a certain action, moving from a "black box" to a transparent assistant.

What's next for A "Smart Convalescence Agent" for post-surgery patients.

Predictive Wound Analysis: Integrating computer vision to allow patients to take photos of surgical sites for early detection of inflammation.

Direct Hospital Integration: Partnering with healthcare providers to feed AuraCare data directly into hospital Dashboards.

Voice-First Interface: Adding a voice agent so patients can report symptoms hands-free during the early stages of recovery.

Built With

• agentic
• ai
• and
• esp32
• flutter
• java
• laravel
• mqtt
• postgresql