PreTra

PreTra: A Webcam-Powered Breathing Garden

This is a submission for the Presage SmartSpectra SDK Challenge.

What I Built

PreTra is a Windows desktop experience that pairs live webcam-based pulse and breathing estimates with a calm, interactive visual environment. It includes a live camera preview, signal-quality guidance, session summaries, and local session history with CSV export.

Each recording receives its own session number. It identifies a recording, not a person, and does not link the same person across sessions.

Demo

[Watch the PreTra live demo]
https://youtu.be/c1FTmZa12TM

The demo shows a consenting participant, the live camera preview, and SmartSpectra readings as they become available.

Code

PreTra source code and setup instructions

The README explains the Windows Flutter setup and how to provide the licensed SmartSpectra runtime files and a local API key. The API key and licensed runtime files are not included in the repository.

How It Responds to Human Signals

PreTra uses SmartSpectra pulse and breathing-rate estimates from a standard webcam. It displays a rate when SmartSpectra reports a stable sample. If lighting, framing, or movement affects signal quality, PreTra provides guidance so the participant can adjust and continue.

I integrated the Windows SmartSpectra C API through Dart FFI. Native metric callbacks are decoded in Dart and streamed to the session state and UI. Stable samples are recorded in the session history and can be exported as CSV.

The displayed confidence is the value reported by the SDK; it is not measurement accuracy. PreTra does not claim 100% accuracy.

Prize Categories

Most Creative or Delightful Use: PreTra presents camera-derived signals within a visually expressive breathing-garden experience.

Best Thoughtful Real-World Use: PreTra is designed for voluntary wellness and personal reflection, with an explicit webcam-consent step, live signal-quality guidance, and local session history.

Consent and Data Handling

Before a session starts, the participant must agree to start a webcam session. Camera frames are used during live processing and are not saved as video.

PreTra stores session summaries and stable signal samples in an unencrypted SQLite database under the current Windows user profile. Records remain there until deleted through Session History. CSV exports are saved under Documents and must be deleted separately. PreTra does not upload or share the local database. During migration from an older version, a local JSON backup may also be retained. SmartSpectra telemetry is disabled in the app configuration. No optional AI service is used in the app.