Finsight

• 

  this shows a pie chart and trend chart of expenditure

• 

  the list of expenses, adding new transactions, edit and delete

• 

  recommendations

The Problem

Most working adults have no clear picture of where their money goes. Spending accumulates across categories without any pattern recognition or guidance — and without a financial advisor, decisions
stay reactive instead of intentional.

What I Built

FinSight is a personal finance app that lets you log your daily expenses, visualize your spending patterns, and receive AI-generated recommendations on where to cut costs and catch relevant deals —
all in one clean dashboard.

How It Works

Transactions are entered manually and stored locally in SQLite. The Dashboard shows a real-time pie chart of your current month's spending by category, plus a trend line chart across all months. The
Notifications tab sends your category-level spending summary to Claude (Anthropic's AI), which returns personalized cut suggestions and relevant deals in a professional financial advisor tone.

Tech Stack

Built with Python, Streamlit, SQLite, Plotly Express, and the Claude API (claude-sonnet-4-6).

What I Learned

This was my first full greenfield build using an AI coding agent. I learned how to structure a spec before writing a single line of code, how to work with Claude as a genuine coding partner rather
than just a code generator, and how production AI features a

Built With

• claude-api
• express.js
• plotly
• python
• sqlite
• streamlit