NeuroNudge

Chromium extension that detects focus drift and restores flow with gentle nudges, while giving parents and students digital wellbeing tools for balanced and mindful screen time.

Comment

Inspiration

NeuroNudge began with a simple frustration shared by nearly everyone on our team. During late-night hackathons, one person would say "just a quick tab" to open YouTube, someone else would check Discord, and within minutes, the entire focus rhythm was gone. Timers and site blockers never solved the real problem. They punished behavior instead of understanding it. We realized what we needed was not control but awareness; a companion that could sense when focus started to fade and help bring it back in a natural way.
We also saw a growing problem around us. Students spend most of their school hours on Chromebooks, professionals work across endless browser tabs, and families struggle to balance work, study, and rest in the same digital space. Everyone needed help managing attention without losing freedom. That is how NeuroNudge was born, as a digital wellbeing tool that guides attention, supports rest, and creates balance across work and life.

What it does

NeuroNudge is a Chromium-based browser extension that observes how you interact with your digital world. It tracks small signals such as tab switches, idle moments, and typing rhythm to understand your state of focus. When it detects that you are starting to drift, it gently reminds you to pause or refocus.
It features a live dashboard that shows how time is spent across sites, labeling them as productive, neutral, or distracting. A scorecard highlights focus streaks, breaks, and recovery times throughout the day. The system predicts drift using behavior patterns and adapts its nudges to your personal rhythm.
NeuroNudge is also designed for families and classrooms. Parents and educators can set time limits, block harmful or distracting sites, and track balance across study and recreation. Instead of rigid control, the system creates flexible digital guardrails that encourage self-regulation and awareness. It transforms digital wellbeing from a rule into a habit.

How we built it

We built NeuroNudge entirely on Chrome Manifest V3 for stability, performance, and privacy. The extension’s core runs through a background service worker that continuously listens for changes in browser activity. It collects and classifies data such as active tab URLs, typing frequency, and idle durations, and uses these to calculate attention metrics.
All data is processed and stored locally using chrome.storage, never leaving the device. The focus engine operates on lightweight JavaScript functions that assign values to productive and distracting states. These signals are aggregated and visualized in the pop-up dashboard through responsive HTML, CSS, and vanilla JavaScript.
The nudge system is powered by conditional triggers. When certain thresholds are met; for instance, spending too long on a flagged site or showing low typing activity, the system initiates a contextual nudge through a pop-up or notification.
A content script runs on each open tab, capturing typing intervals without recording any actual text. This design ensures complete privacy while maintaining meaningful analytics.
For predictive modeling, we used an early version of agent-based logic inspired by Amazon Nova’s ACT framework. This allows NeuroNudge to anticipate when focus is likely to drop and prepare a short intervention in advance.

Challenges we ran into

We faced a handful of real-world hurdles. Service workers nap in MV3, so we cached state, kept wake ups predictable, and avoided missed ticks. Privacy was non negotiable, so signals stay on device and we never record content. Nudge tone matters, so the copy feels like a guide, not a supervisor. Parents need control while users want autonomy, so we built a permission layer that adapts to the mode. On the plumbing side, ports were busy, Python versions did not match Nova, and Playwright needed a full install; we fixed that with a clean Python 3.11 venv, a single bridge port, and a soft fail path. Issue to fix, fast. Worker sleep caused lost context; we added caching, cadence, and refresh-on-activate. Messages failed from the wrong console; we tested from the worker or a page with the content script and added a status pill to prove it is live. Privacy mode muted signals; we show a clear off state. The banner only shows on Productive sites after cooldown when TTD is low; the status pill explains the current state and notifications only trigger on signal change. The Nova bridge needed the right Python, Playwright browsers, and a stable port; we rebuilt the venv, installed the SDK and browsers, and documented quick lsof and kill steps. Re-entry failed on auth pages; we validated URLs, tailored prompts for Google Docs, and wrapped Nova calls so the app keeps working if automation is skipped. We resolved Git conflicts by rebasing, keeping both parent mode and RBF, and pushing a clean main. The best moment was when the banner lined up with the forecast, we clicked Start reset, the terminal showed the calls, Nova ran the break, and we landed back in the doc with the note ready. That was a real yes. The flow now feels simple for users while the system handles the tricky parts quietly in the background.

Accomplishments that we are proud of

We turned a complex problem, digital distraction, into a tool that actually feels supportive. NeuroNudge successfully detects focus patterns without violating privacy, and the interventions feel human and respectful.
The momentum board became a centerpiece of our design. It motivates users by showing small wins rather than enforcing strict quotas. The parental and wellbeing controls extend this philosophy to families and classrooms, allowing both freedom and structure.
We are proud that everything runs fully on-device. No cloud processing, no analytics services, and no data sharing. Every signal is processed locally in the browser, keeping user trust at the core of the product.

What we learned

We learned that attention is not something to control but something to guide. Designing for focus means designing for balance, between effort and rest, autonomy and accountability, technology and wellbeing.
We discovered that real productivity tools do not need to track every click or second. They need to understand the rhythm behind those actions. Small, well-timed feedback loops can make a bigger difference than any timer or blocker.
We also learned that integrating digital wellbeing with parental guidance is not about restriction but collaboration. The most effective systems encourage awareness and self-regulation rather than control.
From a technical standpoint, we deepened our understanding of event-driven architectures, Chrome’s storage APIs, and asynchronous state management. Every part of the project taught us how to turn abstract behavioral models into responsive, human-centered design.

What is next for NeuroNudge

We are developing a web-based companion platform where users can visualize their long-term focus data, explore weekly flow patterns, and earn progress milestones.
The next major feature is a predictive model built on agent-based learning using Amazon Nova ACT. It will allow NeuroNudge to forecast focus drift before it happens and suggest breaks or environment changes to prevent burnout.
We plan to integrate with productivity tools like Notion, Google Workspace, and Linear, allowing the extension to connect focus sessions directly with active goals and projects.
For families and schools, we will expand the wellbeing suite with group dashboards, adaptive limits, and shared progress boards that promote healthy screen habits collaboratively.
Our vision is for NeuroNudge to become a universal digital wellbeing companion: one that senses focus, respects privacy, encourages balance, and helps people reclaim time for what truly matters.

\( E_{\text{focus}} = \sum_{i=1}^{n} \text{flow}i - \sum{j=1}^{m} \text{drift}_j \)

When that energy dips, NeuroNudge helps restore it with clarity and calm.

Built With

Share this project:

Updates