MyQuietQuote
A mortgage lead marketplace where borrowers compare lender offers without ever exposing their real phone number or email.
Most mortgage comparison sites sell borrower data to dozens of lenders, resulting in relentless cold calls. MyQuietQuote inverts that — borrowers keep full control, and all communication runs through a privacy layer.
The Challenge
Two parties needed to negotiate a mortgage without either seeing the other's real contact details — while still feeling like a normal email or text conversation. And lender spam had to be impossible by design, not discouraged by policy.
Approach
I designed the system around a privacy-first architecture: every conversation gets its own disposable proxy identity, and rate limits are structural constraints in code — not moderation rules a human has to enforce. The focus was on making privacy the default, not an opt-in.
Solution
I built a three-portal system — one for borrowers, one for lenders, and one for admins — using Next.js on the frontend and Express on the backend, backed by PostgreSQL for structured data and Redis for real-time message routing and proxy state. The key engineering challenge was making a proxy layer that felt invisible: both sides needed to use their normal email inbox or phone, never a custom app, and still have messages arrive instantly with no visible intermediary. Every architectural decision revolved around that constraint.
Masked communication
Every conversation gets a unique disposable email address and phone number, generated on the fly when a lender is matched. Incoming emails and SMS hit the platform's webhook endpoints, where they are parsed — signatures, quoted text, and headers stripped — then routed to the other party through a fresh proxy address. Neither side ever learns the other's real details, yet the conversation reads like a normal thread.
Matching & protection
A scoring engine matches lenders to applications based on location, loan type, and current capacity. But the real safeguard is a structural rate limiter: lenders get a hard cap on unanswered messages, and the counter resets only when the borrower replies. A lender being ignored simply cannot keep messaging — spam is impossible at the architecture level, not policed after the fact.
Three-portal experience
Borrowers see a clean comparison dashboard where all offers line up side-by-side with rates, terms, and lender response times. Lenders get a pipeline view of active leads with messaging status. Admins have full visibility into message volumes, match quality, and lender compliance — plus the ability to suspend or flag accounts instantly.
Results
Shipped in ten weeks with all three portals production-ready. No borrower contact details have ever reached a lender. Spam is blocked structurally in code, with zero manual moderation needed. Offer comparison collapsed from a multi-week phone marathon to a single session.
Architecture
Have a similar project in mind?
Let's talk about how to bring it to life, end to end.
Explore more work
Serverless backend powering card payments, instalment loans, digital wallets, and automated seller payouts for an Australian social commerce platform.
Read case studyMulti-tenant SaaS for accounting firms — four user tiers, automatic bank syncing, and instant three-way revenue splitting.
Read case study