Built and shipped NextClean in 3 weeks: a full-stack marketplace connecting users with trusted independent cleaners across the UK. The platform handles everything from discovery to payment, with a recommendation engine that matches users to cleaners based on location, availability, ratings, and service history.

The client launched immediately after delivery and onboarded 100+ cleaners with complete Stripe verification within the first 3 weeks of going live.

Booking flow with real-time availability, service selection, and instant confirmation

Cleaner profiles with ratings, reviews, service areas, and verification badges

Stripe Connect integration powering cleaner onboarding, identity verification, and split payments

Admin dashboard for managing cleaners, bookings, payouts, and platform operations

Recommendation engine that matches users to the best available cleaner based on proximity, ratings, and past booking history

Geofenced matchmaking so users only see cleaners who actually serve their area

Review and rating system that feeds back into the recommendation algorithm, improving match quality over time

The core technical challenge was building a recommendation system that combined geofencing with matchmaking logic. It wasn't enough to just show nearby cleaners; the system needed to rank them by relevance using multiple weighted signals: distance, availability windows, historical ratings, repeat booking patterns, and service specialization.

Geofencing added complexity because cleaner service areas aren't simple radius circles. They're irregular zones that overlap, shift, and need to be queryable in real time as users search. We built a spatial indexing layer that could resolve location-to-cleaner matches fast enough to feel instant in the booking flow.

Next.js for the full-stack application (SSR + API routes)

Stripe Connect for cleaner onboarding, identity verification, and marketplace payments

Vercel for deployment and edge performance

Geospatial indexing for location-based cleaner matching

Custom recommendation engine combining proximity, ratings, and booking history

3-week build from kickoff to production launch

100+ cleaners onboarded with full Stripe identity verification in the first 3 weeks

Live and taking bookings immediately after delivery

Recommendation engine actively improving match quality as booking and review data accumulates

Stripe Connect from day one. We chose Stripe Connect's full onboarding flow (including identity verification) rather than building a simpler payment layer. This added upfront complexity but meant every cleaner on the platform was fully verified before their first booking, which was critical for user trust.

Geofencing over simple radius search. A basic "cleaners within X miles" approach would have been faster to build but would have produced poor matches in dense urban areas where service zones overlap. The geofenced approach cost an extra few days but made the matching noticeably better.

Recommendation engine that learns. Rather than static sorting, the algorithm weights shift as more data comes in. Early on it leans heavily on proximity and availability; as reviews accumulate, quality signals get more influence. This means the platform gets smarter the longer it runs.