Personal Subscription Dashboard

The problem

I knew my monthly total. What I didn't know was the real number. The cumulative spend since day one on each service, which tools were quietly duplicating each other, and which ones I hadn't actually touched in months. The information existed, scattered across bank statements and signup confirmations. There was no single place that made it legible.

What I want to build

A personal subscription dashboard that treats financial clarity as a design problem. Not a budgeting app — something narrower and more honest: a tool that shows you exactly what you're paying for, whether it's worth it, and when to stop.

The decisions I'm designing around

• Cumulative spend over monthly cost. Seeing €21.42/month feels manageable. Seeing €514 since you first subscribed to something you barely use doesn't. The real number changes the conversation.
• Duplicate detection by category. Not just a list — an overlap layer that flags when two services cover the same ground. Running Claude Pro, OpenAI, and Cursor simultaneously shows up as a pattern, not three separate line items.
• AI-powered recommendations in context. The analysis layer shouldn't just see costs — it needs the full picture: usage, category, overlap, role. The output should be a direct call: keep, review, or cancel. Specific reasons, no hedging.
• API-connected usage data where possible. OpenAI's usage API tells you whether you're actually using what you're paying for. Recommendations grounded in reality hit differently than ones based on assumptions.
• Cancellation reminders. Passive renewal is the business model. A well-timed reminder breaks that default.

Why it's worth building

This is a small project, but the kind of problem I find interesting: a clear gap, a sharp point of view about what information actually changes behaviour, and a constraint — build something useful without overbuilding it. The "brutal honesty" framing isn't aesthetic. It's a product decision. If the tool softens the numbers, it doesn't work.