Inspiration Procrastination and attention drift are huge unsolved problems for students. Most tools give feedback after performance drops, not during the moment focus is lost. Tether was inspired by a simple idea: intervene in real time, not in hindsight. We wanted to build something that feels like a live decision-support layer for learning and deep work, especially for high school and college students who switch tabs, lose momentum, and struggle to recover quickly.
What it does Tether is a context-aware real-time intervention system that runs in the browser and detects focus breakdown patterns as they happen. It tracks behavior signals such as: keystrokes inactivity duration tab switches interaction activity in the active tab When risk patterns are detected (inactivity, distraction, interruption loops), Tether triggers actionable interventions like: lock-in focus sprint resume task prompt recovery guidance It also includes a Grades & Risk module to connect behavior patterns with academic risk, helping users understand consequences and prioritize next steps.
How we built it We built Tether as a full-stack web + extension system. Core stack: Frontend: React + Vite Backend logic: Node.js + extension service worker logic Browser layer: Chrome Extension (Manifest V3) Real-time state: per-tab session state + event-driven updates Architecture highlights: content scripts capture on-page activity signals background service worker manages per-tab state and active-tab authority tab-safe intervention routing ensures alerts only appear in the currently active tab intervention UI includes persistent warnings, focus actions, and timeline logging web dashboard adds Grades & Risk intelligence and recovery planning
Challenges we ran into Preventing false or duplicate alerts across multiple tabs. Making inactivity detection strict and reliable without noisy triggers. Keeping extension behavior stable across tab switches, refreshes, and reinjection. Synchronizing power state between the web app and extension so “OFF” truly means everything stops. Balancing urgency in alerts without making the experience annoying. Accomplishments that we're proud of Built a working active-tab-only intervention engine. Implemented strict inactivity detection with persistent actionable popups. Added interruption-pattern detection (stop → resume → stop). Created a clean risk/intelligence layer that links behavior to academic outcomes. Delivered a polished, demo-ready system that feels like a real product, not a static prototype.
What we learned Real-time behavior systems are mostly about state correctness, not just UI. Per-tab architecture is essential for trust in browser interventions. Small UX details (timing, wording, dismissal behavior) heavily impact user acceptance. Shipping quickly for hackathons still requires disciplined debugging and event modeling.
What's next for Tether Publish Tether as a production-ready Chrome extension with onboarding. Add optional account sync and long-term focus analytics. Improve context understanding for more websites and workflows. Add adaptive intervention personalization based on user response patterns. Expand from intervention into a full focus operating system with weekly coaching insights.
Log in or sign up for Devpost to join the conversation.