Tinder-style swipe app for finding local service providers. iOS native, real payments, real users — shipped to TestFlight in three to four weeks.
“If you need a dog walker, a handyman, a wedding photographer, or a tutor in your zip code, your options today are: Facebook groups full of scams, Nextdoor full of strangers, Yelp full of paid reviews, or Craigslist (please don't). I built [YourName]Swipe — a Tinder-style swipe app for finding [your vertical] in [your city]. Swipe right, match, message, book, pay, review.iOS native. Real providers. Real bookings.”
Facebook groups, Nextdoor, Yelp, Craigslist, Thumbtack — every existing option is bad UX for service discovery. The Hinge mechanic for finding a contractor is a $0 → $1B idea sitting on the table.
Swiping is visceral. In 60 seconds your audience sees the gesture, hears the haptic tap, watches the match animation, and gets the entire concept. No slides needed.
Auth, realtime chat, AI matching, payments, geolocation, push, image handling, App Store distribution. A portfolio piece most senior engineers don't have.
Don't try to be the universal services app. Pick one service in one metro and own it. The vertical you pick determines the providers you'll seed, the screenshots you'll show, and the Demo Day pitch. Your worked example below is wedding photographers in a major metro — $1.5K–$5K per booking, high-emotion, portfolio-first, a perfect fit for a swipe-card stack.
Don't pick: lawyers, doctors, financial advisors (regulated, slow). Don't pick: plumbers, electricians (emergency-driven, bad fit for swipe). Don't pick: anything < $20 avg transaction.
Email + password via Supabase Auth. Seeker or Provider picker at first signup.
Photo upload, bio, services (max 3), price range, zip + radius.
Stack of provider cards. Right to like, left to skip. Haptics, checkmark, X.
Seeker + provider area overlap → match row. Celebratory modal.
Supabase Realtime channel per match. Text only for MVP.
Date + service + price proposal. Provider accepts or declines.
Stripe Connect. Seeker pays in-app. 90% provider, 10% platform.
Post-booking 1–5 stars + text. Visible on provider's card.
Acceptance test — a real friend, on a real iPhone, completes the full flow from signup to review without your help.
A dev shop will quote you $25,000 for the same scope. That delta is the thesis.
Empty but wired. App on your phone, Supabase project, auth end-to-end.
One-pager: vertical, city, why this vertical, three real local providers you'll invite.
Empty SwiftUI app with TabView: Discover, Matches, Messages, Profile. MVVM scaffold.
Schema deployed. Swift SDK integrated. Signup writes a row you can see in the dashboard.
Sign up / log in / log out. Seeker-or-Provider picker. Session persists across launches.
The swipe deck. The show-off week. End of week, the app looks like the product.
Photo upload to Supabase Storage. Name, bio, services, price, zip, radius.
Tinder-style stack. SwiftUI gestures only — no libs. Haptics, rotation, opacity, checkmark-X overlays.
Cards load from Supabase, filtered by vertical and seeker's zip + radius. Swipes persist.
Area overlap → match row. Celebratory match-screen modal. Matches tab lists them.
Toy becomes marketplace. Chat, booking, real money, reviews. MVP shipped.
Supabase Realtime channel per match. Text messages, instant both sides.
In-chat date + service picker. Provider gets accept/decline card.
Express onboarding, KYC, payment sheet. 90/10 split. Budget 2× time. Test mode only.
Post-booking 1–5 stars + text. Rating average updates on provider's card.
Optional sprint. Where the ambitious students separate themselves.
Claude embeds provider bios and ranks the deck by semantic match to the seeker's stated need.
APNs cert generated. Push fires on match + message + booking status change.
Archive, sign, upload to App Store Connect. Distribute to 5 real testers.
Live build, 30-second pitch, 5 seeded providers, 1 external tester, 2-minute flow. See below.
Stripe Identity check + ID upload to Supabase Storage.
Provider sets availability — seeker only sees open slots.
Provider sends a photo of the finished work in-thread.
3–6 swipeable shots. Critical for wedding photographers.
Drop-down switches between dog walkers, cleaners, trainers.
Embed bios and rank by “big dogs, reactive okay, outdoor natural light.”
Not simulator. You hand someone a phone and they use the app.
Every word earns its place. Practice until it's muscle memory.
Real names, real photos, with permission. No lorem ipsum.
Someone outside the cohort who actually used the app. Tell their story.
Open → swipe 3 → match → chat → book → payment screen.
Paste the full Xcode error into Claude. Don't try to debug in the Apple portal — you'll lose the will to live.
Use test mode end-to-end. Fresh test account per iteration. Symptom is payouts that never arrive.
Restart Xcode (fixes 80%). Then check @StateObject vs @ObservedObject.
Don't. iOS only. Fork to Android after Demo Day if you have customers asking.
The MVP list is sacred. Stretch goals live in Week 4. Never in Week 2.
If you can't recruit five real locals for a free launch spot, you don't have a business — you have a class project.
12 lessons. 3–4 weeks. ~$110 out of pocket. A working iOS app on TestFlight with real providers, real payments, and a story that writes itself.
Enroll in AI Builders Club →Pre-reqs: Modules 1–3 complete · Cohort Slack channel: #cap-serviceswipe