SentriOS

Digital Study Sanctuary - A standalone OS-type app within windows which eliminates all digital distractions that plague the students focus while studying.

Comment

Inspiration

Whenever me and my friends would sit down to study using our laptops, we would get distracted and click on random youtube videos or online games. So I decided to develop an app which silences all digital distractions and gamifies focus boosting and study motivation. I call it - SentriOS.

What it does

SentriOS is a desktop study companion that helps students stay on-task during focused sessions. It combines:

  • Session planning — set your subject, topic, and study duration before each session
  • Live focus tracking — uses the laptop camera to detect presence, attention, and low-light conditions, updating a real-time focus score throughout the session
  • Adaptive break flow — schedules breaks dynamically based on your focus score, with gamified break activities and streak shields
  • AI-powered study support — contextual chat help, end-of-session recaps, and quiz generation, all powered by Amazon Nova via AWS Bedrock
  • Smart browser guard — evaluates whether the pages you browse during a session are relevant to your study goal, using Nova to classify and block distractions
  • Parent accountability — OTP-based parent flow and a cloud-hosted dashboard for monitoring study activity
  • AWS present -- but currently disabled for ease of testing.

gi

How we built it

SentriOS is a local-first desktop app built with Flet for the UI and Python for the core logic. The system is modular across several layers:

  • Session engine — manages the full session lifecycle: start, pause, break, resume, interruption recovery, and completion
  • Camera monitoring pipeline — runs a local computer vision loop using OpenCV and MediaPipe to detect face presence, gaze direction, phone usage, and low-light conditions
  • Focus engine — converts real-time camera signals into a dynamic focus score that rises with attention and falls with distraction events
  • Local SQLite persistence — stores sessions, focus events, streak metrics, and analytics data; app works fully offline
  • Amazon Nova via AWS Bedrock — powers in-session chat, recap generation, quiz generation, and AI-based URL relevance classification
  • AWS cloud layer — DynamoDB for session sync, SNS for OTP delivery, and S3 for hosting the parent dashboard

Challenges we ran into

  • Making the camera and monitoring pipeline reliable across different Windows/Python environments required significant compatibility work, including fallback strategies when MediaPipe API shapes differed between builds
  • Managing desktop UX around the embedded study browser — keeping it positioned correctly, focus-locked to the session window, and properly cleaned up on crashes — required many reliability fixes
  • Ensuring session continuity and crash recovery (detecting orphaned processes, resuming interrupted sessions) was harder than expected on Windows
  • Calibrating the URL guard policy was tricky: we needed Nova to block genuine distractions without frustrating legitimate study workflows like searching for a topic or watching an educational video

Accomplishments that we're proud of

  • Built a complete, working end-to-end product — not just a prototype screen flow
  • Translating real-time behavior signals into a meaningful focus score that students can actually act on
  • Using Amazon Nova in ways that are genuinely tied to studying: session-aware chat, personalized recaps, quiz generation, and distraction detection — not generic AI features bolted on
  • Implementing parent accountability and cloud sync while keeping the core product fully usable offline
  • Turning a personal frustration shared with friends into a structured, measurable study tool

What we learned

  • Real productivity tools live or die by reliability and edge case handling, not just the happy path demo
  • Attention support works best when lightweight AI, behavioral signals, and UX nudges work together — no single layer is enough on its own
  • Local-first architecture is important for privacy, resilience, and a smoother user experience
  • Good distraction control is as much a product design problem as a technical one — the line between helpful enforcement and frustrating restriction is subtle
  • Building for students means respecting flexibility while still holding a firm boundary during focus windows

What's next for SentriOS

  • Personalized distraction rules and coaching that adapt to each student's pattern over time
  • Richer focus analytics and long-term progress insights surfaced directly in the app
  • Expanded parent and mentor views with trend summaries and meaningful alerts
  • Cross-device and classroom-ready deployment options
  • Continuously refining the balance between strict focus protection and student autonomy

Built With

  • amazon-nova-pro-cloud-services:-aws-dynamodb
  • aws-s3-database:-sqlite-(local-first-persistence)
  • aws-sns
  • language:-python-ui-framework:-flet-(desktop-app-ui)-computer-vision-/-ml:-opencv
  • mediapipe-(face/presence-+-distraction-signals)-browser-integration:-pywebview-+-custom-browser-worker-process-logic-backend-sdk:-boto3-(aws-sdk-for-python)-cloud-ai:-amazon-bedrock-foundation-models:-amazon-nova-lite
  • sqlalchemy-orm-data/config-formats:-json
  • yaml-platform/runtime:-windows-desktop-environment-(local-first-app
Share this project:

Updates