What I Built
For the Presage SmartSpectra SDK Challenge, I developed PulseFlow, a desktop application that measures breathing rate using a standard camera and presents the results through an interactive visualisation.
I wanted to explore how contactless physiological sensing could be integrated into a desktop application while keeping the experience simple and understandable.
PulseFlow is built using Node.js, Electron, JavaScript and the Presage SmartSpectra SDK.
Demo
Watch the PulseFlow demonstration on YouTube
The video is presented at 0.6× playback speed. Measurements were originally captured using the live SmartSpectra SDK integration.
How It Works
PulseFlow uses the Presage SmartSpectra SDK to estimate breathing rate from camera input.
The user selects a camera and starts a measurement session. The application then receives measurement and validation information from SmartSpectra.
During the session, the interface displays the current measurement status and provides guidance when the signal is unstable.
Once a stable breathing-rate measurement is available, the application updates the displayed value and uses it to drive a visual waveform and expanding ripple animation.
The animations are generated by PulseFlow using the estimated breathing rate. They are not raw respiratory waveforms returned by the SDK.
Features
- Contactless measurement: Estimates breathing rate using a standard camera.
- Live breathing-rate display: Shows measurements received from SmartSpectra.
- Signal validation: Indicates whether the measurement is stable or unstable.
- Positioning guidance: Displays feedback to help users obtain a better measurement.
- Interactive visualisations: Uses breathing-rate measurements to animate waveforms and expanding ripples.
- Session controls: Allows users to start and stop measurements.
Technology Stack
| Technology | Purpose |
|---|---|
| Node.js | Application logic and SDK integration |
| Electron | Desktop application |
| JavaScript | Application functionality |
| Presage SmartSpectra SDK | Camera-based breathing-rate estimation |
| HTML and CSS | User interface and visualisations |
| Playwright | Automated interface testing |
Development Challenges
One of the main challenges was working with live physiological measurements.
The SDK does not immediately produce a stable reading when a session starts. Camera positioning, lighting and signal quality can affect measurement availability.
I therefore focused on presenting measurement status clearly rather than displaying a value before a sufficiently stable signal was available.
Another challenge was designing meaningful visual feedback. PulseFlow uses the breathing-rate estimate to control its animations, while keeping the distinction between a measured value and an illustrative waveform clear.
I also created automated Electron interface tests using simulated SDK responses to check the main user interactions and application states.
Running PulseFlow
The public repository contains the source code and project documentation:
GitHub: https://github.com/sriramrmenon/pulseflow
To run the application, clone the repository, follow its README installation instructions, configure a valid Presage SmartSpectra API key and connect a compatible camera.
Camera permission must be granted before a measurement can begin.
Consent and Data Handling
PulseFlow is designed around user-initiated measurement sessions.
The user controls when a session starts and stops, and camera access requires the applicable device or operating-system permission.
The application uses camera input for SmartSpectra breathing-rate estimation and displays the resulting measurements and validation feedback in its interface.
Before publishing, verify and complete these disclosures against the implementation and Presage's telemetry documentation:
- Whether camera frames or measurement data are stored, shared, retained or deleted.
- What telemetry or processing is performed by the SmartSpectra SDK.
- Whether any optional external AI services are used.
PulseFlow is an experimental demonstration and is not intended for medical diagnosis, treatment or clinical decision-making.
Prize Categories
I would like to enter PulseFlow in Best Thoughtful Real-World Use, alongside consideration for the overall challenge.
My focus was on making contactless breathing-rate measurement understandable through clear visual feedback, signal-quality guidance and explicit user controls.
What I Learned
Building PulseFlow gave me practical experience integrating a native physiological sensing SDK into an Electron desktop application.
It also reinforced the importance of handling unstable live data, giving users meaningful feedback and communicating what a measurement does and does not represent.
I'm interested in continuing to explore how real-time signals can be used in interactive applications.
Source code: https://github.com/sriramrmenon/pulseflow
Demo video: https://youtu.be/bjeCthulDlg
Top comments (0)