Clarity Browser: Working Smarter When Your Brain Is Overwhelmed

Inspiration

The prompt was deceptively simple: design a product that helps people work smarter when their brain is overwhelmed.

But before we could design anything, we had to define what "overwhelmed" actually means because it gets thrown around a lot without much precision behind it.

We landed on two root causes:

Too many to-do's —> your task list has grown beyond what feels manageable, and instead of motivating you, it paralyzes you
Unmet basic needs —> sleep, food, water, and movement aren't just wellness checkboxes. When they're neglected, your cognitive capacity shrinks

We modeled overwhelm as a function of two compounding variables:

O = f(T, N)

Where $O$ is your overwhelm level, T is task load (volume and complexity), and N is your needs deficit (sleep, nutrition, hydration, movement). Both amplify each other and a neglected body makes a long task list feel impossible, and an impossible task list makes it harder to take care of yourself.

From there, we asked: what would we actually want to use?

The answer came from turning the lens on ourselves. As someone juggling full-time work, multiple active side projects, a deliberate health practice, consistent sleep, regular workouts, and an active social life including networking events, friend time, and dates, the existing tools had a specific gap. A calendar records what's happening. A task list shows what needs to happen. Neither one tells you what to do when nor accounts for the fact that not all hours are equal.

Doing an honest self-evaluation, it became clear that my optimized schedule looks nothing like a standard 9–5. I've been actively learning what my natural rhythms are, and what I needed wasn't another app to log things, I wanted something that will have all my personal and professional tasks together at one place! Thats why we build an electron based browser that became our product, Clarity Browser.

What We Built

Clarity is an Electron-based productivity browser built around cognitive relief. It combines everyday browsing with Jira integration, Google Calendar sync, an adaptive AI coach, and a health-aware notification system — all in a single window.

The core output is a dynamic, personalized schedule — not a static list, but a living plan that accounts for your energy, your task types, and your health in real time.

We defined working smarter as:

$$\text{Smart Work} = \text{Energy Sync} \times \text{Task--Time Match} \times \text{Reduced Task Switching}$$

Energy Sync — your schedule adapts to your health and energy levels, not the other way around
Task–Time Match — the right kind of work gets placed at the right time of day, based on when you do it best
Reduced Task Switching — time blocking protects deep work and reduces the cognitive cost of constant context shifts

When a Jira ticket is due soon, Clarity doesn't just flash a notification. It asks how you're feeling. If you're overwhelmed, it walks you through a progressive, calming flow: one short recommendation at a time, contextual action suggestions, and eventually a full AI coaching conversation if you need deeper support. The AI coach drafts messages to your team lead, breaks tickets into smaller steps, and suggests concrete next actions grounded in the actual ticket content.

Key features include:

A browser-first UI with Arc-style vertical tabs, command palette, and focus mode
Jira two-way sync (fetch issues, transition statuses)
Google Calendar integration with meeting prep and AI-powered reschedule drafts
Adaptive health check-ins that adjust frequency based on sleep, mood, and energy patterns
A morning brief that previews your day and helps you ease into work
A cognitive relief mode that strips away UI noise when you need to focus
A personalization wizard that learns your sleep patterns, focus windows, and work hours

The onboarding collects your sleep schedule, peak focus windows (early morning, morning, afternoon, late afternoon, evening, late evening, night), work hours, and task types — focus work, collaboration, strategy, recovery. Throughout the day, pop-up check-ins track how you're actually doing, capturing energy on a $[0, 100]$ scale and mood via emoji, after sleep, meetings, tasks, and health nudges.

How We Built It

We built Clarity as an Electron app with a React 19 + TypeScript frontend bundled by Vite 8. The UI is styled with Tailwind CSS 4 and animated with Framer Motion. State management uses Zustand stores for tasks, calendar, browser tabs, and energy logs.

The main process handles integrations: a Jira REST client for issue fetching and status transitions, Google OAuth for calendar sync, and an OpenAI integration (GPT-4.1-mini) powering the AI coach across multiple modes — chat, action cards, overwhelm plans, calendar recommendations, and health interventions. Credentials are stored securely using Electron's safeStorage and keytar.

All data is local-first, stored in SQLite via better-sqlite3. The reminder engine calculates due-soon notifications with configurable slots, and the AI relief popup uses a progressive disclosure pattern — surfacing one calming recommendation at a time rather than generating a full plan that itself becomes overwhelming.

Given the 24-hour constraint, we made a deliberate call to split the product into a free/basic tier and a future premium tier. Deeper integrations — Apple Health, Garmin Connect, calendar and email sync — were scoped as paid features. This kept the core demo clean and shippable while preserving the full product vision.

Every AI feature also has a keyword-based fallback so the app remains fully functional without an API key.

Challenges We Faced

Integration complexity was our biggest structural challenge. The version we envisioned pulls live biometric data from Apple Health and Garmin Connect — sleep quality, HRV, activity levels — to inform scheduling in real time. It also integrates with existing calendars and messaging apps so your full context flows in automatically. Getting all of that working in 24 hours wasn't realistic, so we made the scoping call early and protected the core experience.

Getting Vite 8 to play nicely with native modules was a significant technical hurdle. Vite 8 uses Rolldown internally, and getting it to work with vite-plugin-electron, better-sqlite3, and keytar required explicit externalization and rebuilding with @electron/rebuild. The electron module wasn't resolving correctly at runtime until we tracked down the configuration gap.

Designing the AI relief flow to feel genuinely calming required more iteration than expected. Early versions returned large structured plans — which were themselves overwhelming. We had to strip back to one recommendation at a time and carefully tune the prompting to lead with emotional easing before any practical advice. The difference between a tool that helps and one that adds to the noise came down to restraint.

Handling Jira's date-only due dates (no time component) required creative workarounds — appending end-of-day times and adding Jira-specific advance notice windows to make the reminder engine fire at sensible moments.

Fixating on one solution was a softer but real challenge. Once we had a clear direction, it became harder to step back and pressure-test alternatives. Brainstorming breadth before committing to depth would have served us well earlier in the process.

What We Learned

Emotional UX is fundamentally different from feature-driven UX. Building a calming interface required removing things — less text, fewer options, more whitespace. Progressive disclosure isn't just a pattern; it's a philosophy when you're designing for overwhelmed users. The most important design decision we made was choosing what not to show.

Fallback engineering matters as much as the happy path. Real demos fail when APIs are slow or unavailable. Having keyword-matched heuristic plans for every AI call meant we could demo confidently regardless of network conditions — and it made the product genuinely more robust.

Scoping is a product skill. The free vs. premium split wasn't just a time management decision — it forced clarity on what the core value proposition actually was. If you can't describe what the free version does in one sentence, the product isn't focused enough yet.

On the technical side, we gained deep experience with the Electron + Vite 8 stack, native module rebuilding, secure credential storage, and multi-mode AI prompting — chat, action cards, overwhelm flows, and health interventions — all within a single application.

What's Next

Live Jira polling — automatically fire relief notifications when real tickets approach their due dates
Calendar-aware scheduling — use meeting density and energy patterns to suggest optimal focus blocks and warn about overloaded days
Health app integration — connect Apple Health and Garmin Connect to inform scheduling with real biometric data
Team awareness — detect when multiple team members are overwhelmed on the same ticket and suggest collaborative sessions
Browser context integration — use the active tab's content to provide smarter coaching (if you're reading AWS docs while stuck on a timeout error, the coach knows)
Mobile companion — a lightweight app for health check-ins and morning briefs so the wellbeing loop starts before you sit down
Plugin system — let teams add custom integrations (Linear, GitHub Issues, Slack) within the same cognitive relief framework