Inspiration
Mortgage pricing optimization is a p versus np problem. Every lender has their own underwriting model and it's incredibly tricky for a mortgage broker to find the optimal price for a loan across all the lenders on a marketplace given a borrower's unique circumstances and the current standards of the industry (for example setting a specific loan amount will produce a non-optimal price because the models need to conform to the loan amount (this can case a 10-20 bps difference in the end rate).
What it does
It allows you to explore mortgage rates in a similar way to searching for a domain: keep switching the inputs and immediately see results. Users can adjust parameters like purchase price, credit score, loan term, and property details to instantly see how these changes affect available mortgage options from different lenders.
How we built it
We used ClosingWTF's public mortgage APIs to retrieve mortgage rates. We developed a responsive interface that provides real-time feedback as users adjust their inputs, with a focus on speed (our current first rate load time is 1.72s) to create a seamless user experience.
Challenges we ran into
- Optimizing API calls to maintain fast response times with changing user inputs
- Designing a simple yet informative UI that clearly shows the relationship between loan parameters and rates
- Accounting for the complex relationships between different mortgage variables (how changing one parameter affects others)
- Presenting complex financial information in an accessible way that helps users make informed decisions
Accomplishments that we're proud of
- Created a responsive interface that updates mortgage options in real-time as users modify their inputs
- Successfully integrated with ClosingWTF's APIs to display accurate, up-to-date mortgage rates
- Developed a sorting system that makes it easy to compare options based on rate, origination costs, or monthly payment
- Built a tool that simplifies the complex mortgage shopping process for both consumers and brokers
- Achieved sub-2-second initial load time despite complex calculations and multiple API calls
Built With
- closingwtf
- loansifter
- nextjs
- react

Log in or sign up for Devpost to join the conversation.