Old draft — do not submit

Inspiration

A student's workload is scattered across calendars, course schedules, assignments, and exams. A long checklist shows what exists, but it does not make the pileup easy to understand or help someone start. We built Boilerplate around a simple visual metaphor: see the pressure building, then release it by making progress.

What it does

Boilerplate brings coursework and calendars into a student dashboard with an animated boiler gauge. The gauge summarizes upcoming meetings, unfinished assignments, and exam preparation in four zones: Cool, Working, Hot, and Redline. Users can inspect the factors behind the score and adjust event weights. It is a transparent planning estimate, not a measurement of personal stress.

Pip, our draggable companion powered by the Gemini API, helps students identify what is due, choose a manageable next step, and plan study time around their availability. Pip uses current open assignments and suggested free windows rather than guessing deadlines. Study proposals require the student's acceptance before the paired browser extension schedules focus reminders or website blocking.

The focus extension blocks distracting websites during a study session. A calendar view makes deadlines easier to scan, and teams can coordinate around workload with per-team visibility controls. A new member's boiler is hidden by default; the user can choose to share with leads or the whole team.

How we built it

We built the web app with Next.js, React, TypeScript, Tailwind CSS, and Framer Motion. Supabase provides authentication, PostgreSQL storage, and row-level access controls. Calendar and coursework connections bring source data into a shared workload model. The app runs on Cloudflare Workers through OpenNext, with GitHub-based deployment.

The pressure engine combines event duration, event weights, proximity, and unfinished task load into a bounded score. Assignments and major exams use different planning windows so exam preparation can appear before the final week. Completing a task removes that task's contribution without discounting unrelated unfinished work.

Pip calls Google's Gemini API on the server. Each question refreshes the authenticated user's assignment context. Separate request and reply checks keep the assistant focused on planning, and generated study proposals are validated against task ownership and calendar conflicts. Calendar event titles and credentials are excluded from Gemini's availability context.

Challenges we worked through

The hardest part was making a simple gauge reflect real workload honestly. We refined exam preparation, overdue work, recurring events, and completion behavior so the score remains explainable. We also worked through partial calendar coverage, source failures, privacy boundaries between teammates, and keeping the app and extension in sync.

Building an animated companion introduced another set of details: dragging, keyboard controls, responsive positioning, reduced motion, and preserving the conversation as a user moves between pages. Deploying the full application on Cloudflare also required attention to runtime limits and background work.

Accomplishments and learning

We connected the full loop from seeing a workload to choosing a task, planning a study window, and protecting that time. We are proud of the boiler metaphor and Pip's personality, but also of the less visible work: explicit sharing controls, an inspectable formula, server-side ownership checks, and clear handling of missing data.

The project taught us how much trust depends on small details: an overdue task staying visible, a completed assignment disappearing from context, and an assistant acknowledging incomplete availability. We used unit, browser, and workflow checks while iterating. AI assistance, including Codex, supported development and testing.

What's next

We want to validate the workload weights with more student schedules, improve import coverage, and refine how students and teams review their upcoming workload. We also want to keep improving Pip's planning suggestions and the reliability of focus sessions.

Additional tracks

Best UI/UX Hack; Best Use of Gemini API.

Try Boilerplate: https://boilerhub.work/ Source code: https://github.com/dsdsouza/Boiler-hub

Built With

• cloudflare
• gemini
• next.js
• postgresql
• react
• supabase
• tailwindcss
• typescript