AtollCare: Turning Every Phone into the First Link in Emergency Care with Presage SmartSpectra

Toppa — Sat, 14 Nov 2026 07:42:38 +0000

This is a submission for the Presage SmartSpectra SDK Challenge

What I Built

AtollCare turns the smartphone everyone already has into the first monitoring point in an emergency-care journey, connecting families, emergency crews, and hospitals through one continuous patient record.

In emergencies, critical patient information is often lost between the initial call for help, the arrival of an emergency crew, and hospital admission. Families may struggle to communicate what happened, emergency crews need to collect information quickly, and hospitals often receive incomplete handoff information.

AtollCare addresses this problem through three connected experiences:

Family's Phone (/pre-arrival): Families scan a health centre QR code, receive the instruction to call 102 first, provide consent, and record a 30-second video explaining what happened. Presage SmartSpectra measures heart rate and breathing rate through the phone camera in real time. ElevenLabs Scribe transcribes the family's account, while Gemini examines selected video frames for visible signs and provides frame-based evidence. Nothing is shared with the care team until the family presses Send.
Emergency Crew (/crew): The crew receives the incoming patient packet and can initiate the transfer with one tap. A positioned smartphone continues monitoring using Presage, while crew members add SpO₂, blood pressure, temperature, and consciousness measurements. Spoken clinical notes are converted into structured handoff information, including medication names and dosages. ElevenLabs reads hospital questions aloud.
Hospital (/hospital): The receiving hospital sees the latest Presage readings, trends, measurement timestamps, the family's original account, and a Gemini-generated SBAR summary. Every measurement includes its source and age, helping clinicians understand when and where the information was collected. The hospital can accept the handoff and specify what should be prepared.
Network Insights (/insights): AtollCare provides network-level insights into acceptance times, journey durations, and alerts across connected facilities, powered by Tiger Data.

The central idea is simple: Capture the information once, preserve its context, and make it available to everyone responsible for the patient's care.

Demo

youtube video

The demo showcases the complete AtollCare journey, from family video recording and live Presage signal monitoring to emergency crew handoff, hospital acceptance, and network insights.

Code

github repo

Technology stack:

Next.js 16 and React 19 for the frontend.
FastAPI and Pydantic for the backend engine.
Presage SmartSpectra with @smartspectra/node-sdk for physiological signal monitoring.
ElevenLabs Scribe for speech-to-text.
ElevenLabs Text-to-Speech for voice interaction.
Gemini for video-frame analysis, identification reading, structured SBAR generation, and medication extraction.
Tiger Cloud with TimescaleDB 2.31 for time-series storage and analytics.

The backend generates types from its OpenAPI specification to maintain consistency between the frontend and backend.

How It Responds to Human Signals

Presage SmartSpectra is at the heart of AtollCare's monitoring experience. It enables an ordinary smartphone camera to contribute physiological information to the emergency handoff without requiring a dedicated monitoring device for the initial assessment.

I integrated the SmartSpectra SDK through a dedicated Node.js service running @smartspectra/node-sdk.

The phone captures camera frames and streams them over HTTPS to the Node service. SmartSpectra processes the camera input and returns framing hints, heart rate, breathing rate, and associated waveforms.

AtollCare uses Presage's own signal-stability verdict to determine whether a reading is suitable for scoring. Readings outside the supported FDA-cleared measurement ranges are not treated as valid scored measurements; instead, the interface directs users to obtain a manual measurement.

The application responds to the following signals:

Heart rate: Displays the latest available reading and its trend.
Breathing rate: Provides additional physiological information during monitoring.
Waveforms: Help visualize the underlying signal output.
Signal stability: Determines whether readings are sufficiently reliable to be included in the scored record.

One important design decision was to avoid treating every returned number as reliable clinical information. AtollCare distinguishes between measurements that pass Presage's stability checks and those requiring manual confirmation. Measurement timestamps and sources are also preserved so hospital staff can understand the context of each reading.

The family's initial video is treated as a bystander estimate and is never scored as a clinical measurement.

Presage is therefore more than a sensor integration. Its signal-stability assessment directly influences how information is presented and whether it can contribute to the patient record.

Prize Categories

Best Thoughtful Real-World Use

AtollCare is primarily entering the Best Thoughtful Real-World Use category.

The inspiration comes from a real-world emergency-care challenge: a patient's story and important observations can become fragmented as responsibility moves between family members, emergency responders, and hospitals.

AtollCare is designed around the people involved in that journey.

For families, it provides a simple way to communicate what happened without requiring medical terminology or technical knowledge.

For emergency crews, it reduces repetitive information collection through speech-based notes, structured measurements, and a shared patient packet.

For hospitals, it provides an organized handoff with measurement sources, timestamps, trends, and the family's original account.

For healthcare networks, it provides insights into acceptance delays and patient journey times across facilities.

What makes AtollCare thoughtful is how it combines live physiological signals with human context. It does not blindly trust AI-generated observations or unstable readings. Presage's stability verdict determines which measurements can be scored, while Gemini's visible-sign observations must reference actual video frames.

The goal is to improve communication, continuity, and situational awareness during emergencies, while keeping healthcare professionals responsible for clinical decisions.

Consent and Data Handling

AtollCare obtains consent before camera recording begins. The family is instructed to call 102 first, reviews the consent information, and explicitly initiates the 30-second recording. Camera frames are transmitted securely over HTTPS for Presage signal processing, while speech and selected video frames are processed by the integrated AI services. The family can review the collected information before pressing Send, and the patient packet is not shared with the emergency crew or hospital until the family explicitly submits it. AtollCare is designed to avoid covert monitoring and does not provide medical diagnoses or replace professional medical decisions.

gemini

Toppa — Thu, 20 Feb 2025 10:47:54 +0000

gemini long text

Toppa — Wed, 13 Nov 2024 19:31:37 +0000

DEV Community: Toppa