This is a submission for the Presage SmartSpectra SDK Challenge
What I Built
Pace is a camera-powered human rhythm companion built with Electron and the Presage SmartSpectra SDK.
The idea behind Pace is simple: what if technology could respond to the subtle rhythms of the human body instead of simply displaying measurements?
Pace transforms real-time physiological signals into a quiet, interactive experience.
With the user’s consent, Pace uses a webcam and Presage SmartSpectra to obtain live pulse and breathing-rate measurements. It presents valid readings and uses the measured breathing rate to drive a responsive visual orb.
Rather than building another dashboard for physiological data, I wanted to explore a more natural way of interacting with technology — one that connects human rhythm with digital experiences.
Pace is designed for general wellness and personal awareness. It is not a medical device.
Demo
Watch the Pace demonstration video
The demonstration showcases the desktop application, camera consent, SmartSpectra integration, live pulse and breathing measurements, and the interactive breathing experience.
Code
GitHub Repository – Musa108/Pace
Pace is built with:
- Electron
- React
- TypeScript
- Presage SmartSpectra Node.js SDK
The repository includes the application source code, architecture documentation, environment configuration example, setup instructions, and tests.
To run the application locally, users need Node.js 20 or newer, a compatible webcam, network access, and their own authorized Presage SmartSpectra API key.
How It Responds to Human Signals
Pace integrates the official Presage SmartSpectra Node.js SDK to access real-time camera-based physiological measurements.
The two primary signals used in the experience are:
- Pulse rate: Displayed in beats per minute when the SDK provides a valid measurement.
- Breathing rate: Displayed in breaths per minute when valid. This signal also drives the visual breathing orb.
The integration works as follows:
- The user opens Pace and chooses to begin a session.
- The user selects a camera, reviews the consent information, and explicitly starts sensing.
- Electron initializes the SmartSpectra Node.js SDK and manages camera access.
- The SDK processes camera input and emits measurement data, validation information, and session states.
- Pace validates incoming measurements and displays readings only when they meet the SDK’s validity requirements.
- The breathing orb responds to valid breathing-rate measurements, connecting the user’s measured rhythm to the visual experience.
A key implementation decision was keeping SDK operations, API credentials, camera management, and measurement validation in Electron’s main process. The React renderer communicates through a restricted preload bridge.
Pace does not generate simulated measurements. The displayed readings and breathing interaction are driven by actual SDK outputs.
Prize Categories
Most Creative or Delightful Use
I was inspired by the idea that technology could become more responsive to the person interacting with it.
Most applications treat physiological measurements as numbers to be monitored. I wanted to experiment with turning those measurements into an experience.
Pace brings together camera-based human sensing, live physiological signals, and a responsive visual breathing companion.
The result is an attempt to make human-computer interaction feel a little more natural, personal, and engaging.
Best Thoughtful Real-World Use
Pace explores how people can interact with their own breathing rhythm through an accessible desktop experience.
It is intended for people interested in personal awareness, breathing-focused interactions, and exploring how technology can respond to human signals.
The experience is designed around voluntary participation. Users explicitly grant camera consent before sensing begins and can stop the session from within the application.
Pace is intended for general wellness and informational use only. It does not diagnose medical conditions, prescribe treatment, or make medical decisions.
Consent and Data Handling
Privacy and user participation are important parts of Pace’s design.
Camera consent and control
Pace does not request camera access automatically when the application launches. Users must choose to begin a session, select a camera, accept the in-app consent checkbox, and start sensing. Users can stop camera capture from within the application.
On-device processing
Raw camera images and video are processed on the user’s device by SmartSpectra. The SDK’s face-derived processing is also performed on-device. Pace does not upload or store camera recordings.
Measurements and third-party processing
An important distinction is that, although camera images and face-derived data remain on-device, SmartSpectra automatically sends derived measurements, including pulse, breathing rate and HRV when available, together with session identifiers and request metadata to Presage for AI Insights.
Presage’s published SDK privacy policy states that these Insights are generated using Anthropic. Therefore, derived measurements are also shared with Anthropic through Presage’s processing.
Pace does not independently call an AI service or supply custom prompt text. However, the SDK’s automatic Insights processing cannot currently be disabled by the application.
Storage and retention
Pace does not persistently store camera recordings or local session history. Its active session summary is held in memory and discarded when the user leaves the session.
Presage stores Insights requests and responses, including measurements, under its published retention policy. These are retained while the developer account remains active and deleted within 60 days after the account closes. Presage also documents separate retention periods for usage records, telemetry, authentication records, and logs.
SDK telemetry
Pace sets enableTelemetry: false to disable optional diagnostic telemetry. This does not disable SDK authentication, model delivery, usage records, or automatic Insights processing.
AI services and analytics
Pace has no separate third-party AI integration or application analytics service. However, Anthropic is involved in the SmartSpectra SDK’s automatic Insights processing, as explained above.
Users should understand this processing before consenting to a session.
For further details, see the Presage SmartSpectra SDK Privacy Policy.
Pace is an experimental general-wellness experience, not a medical device, and its measurements must not be used for diagnosis or treatment.
Agent Session
Pace was developed with assistance from OpenAI Codex.
The agent-assisted development process included application architecture, SmartSpectra SDK integration, Electron security boundaries, measurement validation, session management, and testing.
View the public Codex agent session

