InstaBudget
AI-powered budgeting app that logs expenses from photographed receipts and voice notes.
Problem
Manually logging every purchase is tedious enough that most budgeting apps get abandoned within weeks — the friction of data entry outweighs the value of the insight.
Solution
InstaBudget lowers that friction with three ways to log a transaction: photograph a receipt, describe a purchase out loud, or enter it by hand, with AI handling extraction for the first two.
Technical Approach
A Next.js and React front end talks to Supabase for data storage and authentication.
Receipt photos are parsed with the Google Gemini API to pull out merchant, line items, amount, date, and category; voice recordings are transcribed with OpenAI Whisper before being parsed the same way.
A budgets dashboard tracks category-specific limits and surfaces spending summaries, breakdowns, and threshold alerts.
Implementation
- Built the receipt-scanning pipeline using the Gemini API to extract structured transaction data from photos.
- Added voice-based transaction entry using OpenAI Whisper for transcription.
- Designed the budgeting dashboard, including category limits, spending breakdowns, and progress alerts.
- Set up Supabase for authentication and data persistence.