BidWise

• 

  BidWise

Inspiration

Property research can be hard to piece together, especially when you’re trying to understand a home’s records, possible repairs, and financial tradeoffs. I wanted to make that research easier to follow and help people with limited budgets see what is known, what is uncertain, and what they should check next.

What it does

BidWise helps users research one New Jersey property at a time. Enter an address or paste permitted property-page text, review and confirm the matching parcel, then explore available property records, flood screening, and comparisons between supplied claims and official evidence. Users can also edit renovation assumptions and compare cash resale and rental scenarios.

Sources, assumptions, and unknowns stay visible. The calculations are deterministic, while AI provides a conditional explanation and suggested next steps. BidWise does not guarantee a property’s condition, title, value, or investment returns.

How I built it

I used the Devpost Learn Skill Pack to move from project scope to product requirements, technical specification, and an implementation plan before building. The app combines New Jersey address and parcel data, flood screening, a map, claim extraction from pasted text, and editable financial scenarios. AI helps interpret evidence and estimate some costs; code handles the financial calculations.

What I learned

I learned to make uncertainty part of the product instead of smoothing it over. A missing record is not proof that nothing happened, and a parcel match needs to be checked before sources are combined. I also learned to keep AI explanations tied to visible evidence and to recalculate financial results from explicit assumptions.

Challenges

The hardest part was making inconsistent and incomplete public data useful without implying it was complete. Address and parcel sources can disagree, records may be missing, and AI providers have availability and quality limits. I built confirmation steps, source labels, warnings, and recovery paths around those constraints. Some broader acceptance work, including noisy-page AI evaluation and manual assistive-technology review, remains to be completed.

Built With

• ai
• gemini
• groq
• mapbox
• next.js
• react
• typescript