Subscription Saver

Dashboard
Upload Bills
Detector

Inspiration

Forgotten subscriptions silently drain bank accounts through auto-renewals — what economists call "phantom financial drains". We wanted a smarter tool that treats subscription management as an optimization problem, alerting users before money leaves their wallets.

What it does

Subscription Saver is a web dashboard that tracks, scores, and prunes recurring subscriptions using:

Centralized Analytics — monthly/yearly spend breakdown by category
Predictive Alerts — renewal and free-trial expiry notifications before charges hit
Utility Scoring Engine — flags low-value subscriptions via:

$$U_i = \frac{W_i \cdot F_i}{C_i}$$

where $W_i$ = priority weight, $F_i$ = monthly usage frequency, and $C_i$ = cost. If $U_i < \theta$, the subscription is flagged for cancellation.

How we built it

Built on a modern full-stack ecosystem (GitHub):

Next.js (App Router) + React for SSR and navigation
PostgreSQL via Supabase + Drizzle ORM for type-safe database access
Supabase Auth for secure user sessions
Tailwind CSS + shadcn/ui for UI; Recharts for dynamic spending charts

Challenges we ran into

Recurring interval math — handling weekly, monthly, quarterly, and annual billing across leap years, variable month lengths, and time zones
Balancing $W_i$ vs $F_i$ — a domain name may have low $F_i$ but high $W_i$; tuning the formula to avoid penalizing rare-but-vital services required multiple iterations
Ecosystem version alignment — modern framework changes caused type mismatches in charting libraries, resolved with custom wrapper interfaces

Accomplishments that we're proud of

Zero-lag performance via fast server-side actions and efficient state handling
Glanceable UI — complex date math distilled into instant status badges
Actionable math — $U_i$ turns abstract spend into a clear keep-or-cancel decision

What we learned

Better server/client data partitioning for secure and responsive applications
Stronger multitenant data handling to keep user records isolated
A simple formula like $U_i = \frac{W_i \cdot F_i}{C_i}$ can make digital spending decisions much clearer