An affirming, calm to-do list + life tracker. Built per SPEC.md. Phase 1 is
complete: a fully usable app that runs locally with zero credentials.
- Today screen: daily greeting, check-in entry/summary, a rule-based personalised plan, quick-log buttons, and quick-tick tasks.
- Daily check-in: context-aware flow (core wellbeing + gym/tennis questions on scheduled days), delightful and skippable.
- Areas (categories): data-driven, editable, recolourable, reorderable, with tasks (satisfying tick-off + confetti), session/metric logging, and trends.
- Uni: tasks, upcoming deadlines, study-session logging (the learning-science engine lands in Phase 3).
- Trends/Review: affirmation-first weekly review + per-category charts.
- Settings: areas, weekly schedule, planner-weight tuning, integration status.
- PWA: installs to your home screen and works offline.
npm install
npm run devOpen http://localhost:3000. Default categories/metrics are seeded automatically.
All data is stored in your browser's localStorage — no account needed.
Next.js (App Router) + TypeScript · Tailwind CSS v4 · Recharts · TanStack Query.
The data layer (lib/db) is an interface with two implementations:
LocalDB (default) and SupabaseDB (ready to switch on).
Phase 1 runs local-first. To sync across devices later:
- Create a free project at supabase.com.
- In the SQL editor, run
supabase/schema.sql(creates tables + RLS). - Create
.env.local:NEXT_PUBLIC_SUPABASE_URL=your-project-url NEXT_PUBLIC_SUPABASE_ANON_KEY=your-anon-key NEXT_PUBLIC_DATA_BACKEND=supabase - Restart
npm run dev. The app now reads/writes Supabase (per-user via RLS).
Auth (magic-link) wiring plugs into the same client and is part of finishing the Supabase switch — ask before we set it up, since it touches your credentials.
Not done yet (you chose to skip for now). When ready: push to GitHub, import the repo on vercel.com, add the same env vars, and deploy.
- Phase 2: Google Calendar (OAuth read/write), smarter planner, weekly review.
- Phase 3: OnTrack/Doubtfire sync, time-estimation learning, backwards planning/burn-down, cited study-technique engine.