SplitSmart

Inspiration

Every barkada knows the struggle someone pays for dinner, someone pays for drinks, and at the end of the night nobody knows exactly who owes what. The awkward "utang" conversation ruins the vibe. I built SplitSmart to fix this forever.

What it does

SplitSmart is an AI-powered bill splitter and group expense tracker. Just describe your expenses in plain text and SplitSmart instantly:

• Calculates exactly who owes what to whom
• Shows running balances per person
• Tracks pending vs paid payments with checkboxes
• Generates a shareable summary for your group chat
• Saves all your split history in a searchable dashboard

How I built it

Built entirely using MeDo — no manual coding required. I described the app in natural language and MeDo generated the complete full-stack application including React frontend, Supabase database, and Gemini 2.5 Flash AI for expense calculation. The entire app was built and deployed in under 2 hours.

Challenges I ran into

Getting the AI to accurately parse natural language expense descriptions and compute correct balances was the key challenge. MeDo's AI integration handled this seamlessly through Supabase Edge Functions.

What I learned

Natural language is the most powerful interface for finance apps. People don't want to fill out forms — they just want to describe what happened and get the answer.

What's next

• Multi-currency support
• Recurring expense tracking
• Direct payment integration (GCash, PayMaya)

Built With

• css
• gemini-2.5-flash
• medo
• react
• supabase
• tailwind
• typescript