TracWyse

I discovered I was getting charged monthly for SMS alerts and had no idea. That moment made me realize how many hidden or untracked charges quietly drain our accounts. I wanted a way to track all these costs automatically

It spools your data from your bank SMS sender everyday by 8 am through the cron job set up and organizes your inflows and outflows into categorizes so that you're able to see where your money is going into.

How we built it

Using bolt.new, I built a fullstack web app that:

1. Accepts forwarded SMS via webhook
2. Uses OpenAI to analyze & categorize transactions
3. Stores everything in a Postgres database
4. Lets users view & download monthly spending summaries
5. Runs a daily sync at 8am to update budgets

Challenges we ran into

1. Couldn’t access real SMS data, so I used test data.
2. SMS parsing is hard, they’re often unstructured.
3. Figuring out how to manage month-by-month budgets took time.
4. I needed to keep each user’s data private and separate.

Accomplishments that we're proud of

Used test data for the database and it was able to classify spendings based on different categories which was deployed to Netlify and the documentation is available on Supabase

What I learned

I learned that:

1. Small, frequent charges add up fast.
2. Most people struggle to track their daily spendings.
3. SMS alerts are an untapped data source for personal finance — especially in Nigeria. Technically, I learned how to parse SMS, build an AI-powered categorization system, and generate monthly financial reports.

What's next for TracWise

I'm finishing the SMS integration flow and testing with more real-world cases. The prototype is already functional with test data and shows monthly categorized reports that can be downloaded in PDF, Excel, or CSV formats.

Built With

• css
• java
• python
• typescript