AI Coach

Inspiration

Most budgeting apps are built for adults with salaries and credit cards. Students live a completely different financial reality — irregular allowances, impulsive bubble tea runs, and zero financial education. We wanted to build something that actually speaks that language. The idea came from a simple frustration: why is there no app that tells you "hey, you spent more on snacks than on textbooks this month" and actually makes you laugh about it while helping you do better?

What it does

Track daily spending across 8 categories (food, transport, entertainment, education, shopping, health, snacks, other) Set savings goals with deadlines, custom emojis, and color themes Get AI to auto-categorize your transactions with a single tap Ask your AI coach anything Get roasted by AI for your spending habits (the "Roast Me" feature) Build daily logging streaks and unlock achievements

How we built it

Frontend: React 18 + TypeScript + Vite Styling: Tailwind CSS Auth + Database: Supabase (PostgreSQL with Row Level Security) AI: Groq API running llama-3.3-70b-versatile Icons: Lucide React

Challenges we ran into

Getting the AI to return valid JSON consistently for transaction categorization required tight system prompts with fallback parsing.

Accomplishments that we're proud of

The "Roast Me" feature is genuinely funny and students actually want to use it A fully working AI coach Clean, dark UI that doesn't feel like a boring finance app Gamification (streaks + 9 achievements) that makes daily logging feel rewarding

What we learned

Supabase RLS is powerful but needs to be set up correctly before any data operations work

What's next for CentsAble

Receipt scanning: Take a photo, AI extracts the transaction automatically Weekly email summaries: "Here's how you did this week, and here's your roast"

Built With

• postgresql
• react
• supabase
• typescript
• vite