This is a submission for the Presage SmartSpectra SDK Challenge
What I Built
PulseFlow — Your body, in real time.
I built PulseFlow, a Windows desktop application that transforms an ordinary webcam into a real-time, contactless physiological sensing experience using the Presage SmartSpectra SDK.
The idea started with a simple question: What if exploring physiological signals didn't require a smartwatch, fitness tracker, or specialized hardware?
I wanted to create something that went beyond displaying numbers. My goal was to make human-sensing technology approachable through a minimal desktop interface, live visualizations, and a guided breathing experience.
Real-Time Physiological Monitoring
PulseFlow uses a standard webcam and SmartSpectra to estimate two physiological signals:
- Heart Rate: Pulse rate in beats per minute (BPM).
- Breathing Rate: Respiratory frequency in breaths per minute (br/min).
The application presents these measurements through live numerical indicators and time-series charts.
It also provides camera status, signal-validation feedback, and clear waiting states when measurements are unavailable.
During testing with my webcam, SmartSpectra successfully provided live measurements including 83.7 BPM and 7.2 breaths/min.
These were genuine measurements from a test session, not hardcoded or simulated values.
Guided Breathing Experience
I wanted PulseFlow to offer an interactive experience rather than function only as a physiological monitoring dashboard.
The application includes a guided breathing session with:
- Configurable inhalation and exhalation durations.
- An animated breathing indicator.
- Start, pause, resume, and end controls.
- A synchronized session timer.
- Collection of valid physiological measurements during the exercise.
The breathing animation follows the configured timer. It operates independently of the measured physiological signals and does not fabricate breathing-rate values.
Session Summaries
After a breathing exercise, PulseFlow presents:
- Session duration.
- Average observed heart rate.
- Average observed breathing rate.
- Valid pulse sample count.
- Valid breathing sample count.
- Session completion status.
For example, one completed 60-second session produced:
| Measurement | Observed Result |
|---|---|
| Duration | 60 seconds |
| Average heart rate | 84.9 BPM |
| Average breathing rate | 8.0 breaths/min |
| Valid pulse samples | 69 |
| Valid breathing samples | 15 |
| Status | Completed |
These are results from one demonstration session, not claims about physiological improvement.
Design and Experience
PulseFlow follows a minimal, monochromatic design direction influenced by modern developer tools and Geist-inspired typography.
I focused on large, readable measurements, subtle borders, smooth animations, contextual notifications, and an uncluttered interface.
I wanted to make a complex sensing experience feel simple and understandable.
Demo
I recorded a demonstration of PulseFlow running on Windows with a real webcam and the Presage SmartSpectra SDK.
The video showcases the desktop interface, live heart rate and breathing rate estimates, guided breathing, and the completed session summary.
Watch the PulseFlow demo on YouTube
For privacy, the camera imagery in the public recording is obscured. The displayed physiological measurements come from the live SmartSpectra integration.
The application runs locally as an Electron desktop application. Public web hosting is not required to use it.
Code
The complete source code is available in my public GitHub repository:
The repository contains the desktop application, SmartSpectra integration, measurement-processing logic, breathing-session functionality, interface assets, tests, and setup instructions.
Technology Stack
| Technology | Purpose |
|---|---|
| Electron | Desktop application runtime |
| Node.js | Native application and SDK services |
| TypeScript | Application logic and type safety |
| Presage SmartSpectra SDK | Contactless physiological sensing |
| Protocol Buffers | Measurement decoding |
| HTML & CSS | Interface and animations |
| Canvas API | Camera preview and live charts |
| esbuild | Application bundling |
How to Run PulseFlow
1. Clone the repository
git clone https://github.com/Ezequie1Sc/PulseFlow.git
cd PulseFlow
2. Install the dependencies
npm ci
3. Configure the Presage API key
Create your local environment file from the included example.
On Windows PowerShell:
Copy-Item .env.example .env
Configure the API key in .env:
SMARTSPECTRA_API_KEY=your_presage_api_key
4. Launch the application
npm run dev
You will need a compatible webcam, camera permissions, and a valid Presage SmartSpectra API key.
The repository's README includes additional documentation and troubleshooting guidance.
No separate application backend or database is required.
How It Responds to Human Signals
SmartSpectra is the foundation of PulseFlow.
Rather than simulating physiological data, the application uses live measurements generated from webcam input.
1. Capturing Live Signals
After the user explicitly consents to camera monitoring, PulseFlow initializes the SmartSpectra Node.js SDK and connects to the selected camera.
The application requests these signals:
requestedMetrics: [
MetricType.PULSE_RATE,
MetricType.BREATHING_RATE,
]
SmartSpectra processes camera input and provides physiological measurements together with signal-quality information.
2. Handling SmartSpectra Events
The application listens for SDK events including:
| SDK Event | Application Behavior |
|---|---|
processingStatus |
Updates processing information |
validationStatus |
Provides signal-quality guidance |
metrics |
Receives physiological estimates |
videoOutput |
Provides camera preview frames |
error |
Handles SDK-related errors |
Physiological measurement messages are decoded using the SDK's Protocol Buffers definitions.
3. Processing and Validating Measurements
One of my main technical priorities was avoiding misleading readings.
PulseFlow processes the incoming measurements and filters samples before displaying them.
If a measurement is unavailable or invalid, the interface communicates that state rather than generating an artificial value.
For example, the dashboard can display a waiting state while SmartSpectra attempts to acquire an acceptable signal.
4. Real-Time Responses
When SmartSpectra provides an accepted physiological measurement:
- PulseFlow updates the corresponding heart rate or breathing rate indicator.
- The historical chart incorporates the new reading.
- An active guided breathing session collects the valid sample.
- The accepted measurements contribute to the session summary.
The animated breathing indicator operates independently of the measured respiration signal.
This separation ensures that visual breathing guidance is not mistaken for real physiological data.
5. Integration Challenges
During development, I encountered challenges involving camera framing, frame rate, signal quality, and SDK validation.
For example, SmartSpectra provided feedback such as:
Place more of the chest in view.Use a camera mode that provides at least 25 frames per second.
These messages helped me understand that successfully capturing video does not automatically mean a valid physiological signal is available.
I adjusted the capture conditions and camera positioning until I could obtain live heart rate and breathing rate measurements.
I also worked on separating native SDK processing from the Electron renderer, maintaining a modular architecture for camera handling, measurement processing, and user interface updates.
This experience taught me the importance of testing with real hardware, handling unavailable measurements correctly, and designing interfaces that communicate uncertainty.
Prize Categories
I am entering PulseFlow in both special prize categories.
Most Creative or Delightful Use
The inspiration behind PulseFlow was to make physiological sensing feel like a natural, interactive desktop experience.
Instead of building only a traditional measurement dashboard, I combined real-time heart rate and breathing rate monitoring with guided breathing, subtle animations, live charts, and session summaries.
I wanted the interface to feel calm and approachable while demonstrating the capabilities of contactless human sensing.
PulseFlow brings together native desktop development, real physiological signals, and interaction design in an experience that does not require specialized wearable hardware.
Best Thoughtful Real-World Use
PulseFlow explores a practical use of SmartSpectra: helping individuals observe physiological estimates during voluntary breathing exercises using an ordinary webcam.
The experience is intended for people curious about physiological sensing and general wellness exploration, particularly those who do not have wearable monitoring devices.
I designed PulseFlow around several usability principles:
- Accessibility: A standard webcam is the primary sensing device.
- Consent: Monitoring requires explicit user authorization.
- Transparency: The interface communicates when measurements are unavailable.
- Control: Users can start and stop monitoring.
- Clarity: Live measurements are presented through readable indicators and charts.
- Privacy: The application does not require user accounts or a dedicated physiological-history database.
PulseFlow does not provide medical diagnoses, treatments, or clinical recommendations.
It is an experimental demonstration of contactless sensing and user-controlled interaction.
Consent and Data Handling
Because PulseFlow uses camera-based physiological sensing, consent, transparency, and responsible data handling were important parts of the application design.
Camera Consent
PulseFlow requires explicit user consent before physiological monitoring begins.
The application explains the purpose of camera access and provides controls for starting and stopping monitoring.
The user remains in control of the camera-based experience.
Camera Input and Physiological Data
PulseFlow uses webcam input through the Presage SmartSpectra SDK to obtain physiological estimates.
Camera frames can be displayed in the desktop interface as a local preview.
Accepted measurements are used to update:
- Live physiological indicators.
- Historical charts.
- Guided-session observations.
- Session summaries.
The public demonstration obscures camera imagery to protect the participant's identity.
Storage and Retention
PulseFlow does not implement a dedicated persistent database for physiological measurements, user accounts, or session histories.
The application's measurement-processing functionality is designed around temporary data used during monitoring and guided breathing sessions.
This describes PulseFlow's application-level architecture. It does not establish the independent storage or retention behavior of Presage's SDK or services.
Data Sharing and SDK Telemetry
PulseFlow configures the SmartSpectra SDK with:
enableTelemetry: false
This disables the configurable SDK telemetry option used by PulseFlow.
However, this setting does not necessarily disable SDK authentication requests, required network communication, or any provider-side processing.
The application does not implement an additional backend for uploading physiological session histories.
For details regarding SDK-level data handling, refer to the official Presage documentation.
Optional AI Services
PulseFlow does not require an external generative AI service for physiological measurements or guided breathing.
Physiological estimates originate from the Presage SmartSpectra SDK.
Deletion and User Control
Users can stop monitoring and manage or reset their active breathing session through application controls.
PulseFlow does not provide persistent cloud-based physiological histories.
Temporary measurement information is intended for the active application experience rather than long-term storage.
Responsible Use
PulseFlow is an experimental application intended for education, technical demonstration, and general wellness exploration.
It is not a medical device and must not be used for:
- Medical diagnosis or treatment.
- Clinical decision-making.
- Emergency monitoring.
- Covert surveillance.
- Non-consensual physiological monitoring.
Session Summary Example
One of the final features I implemented was the measurement-based session summary.
After completing a 60-second exercise, PulseFlow displayed the observed heart rate and breathing rate averages, valid sample counts, and completion status.
Building PulseFlow taught me that real-time physiological sensing requires more than successfully connecting an SDK.
Camera conditions, signal validation, measurement availability, native integration, and user experience must work together.
The most valuable lesson was that an honest interface should explain when a measurement is unavailable instead of inventing data.
I am grateful to Presage and DEV Community for the opportunity to explore contactless human sensing through this challenge.
PulseFlow — Your body, in real time.




Top comments (0)