Delusional Leap

The app that turns Gabby Beckford's 1M+ inspired scrollers into daily action-takers. 3 Moves a day. AI-verified. Dreams tracked. Powered by RevenueCat.

Comment

Inspiration

I AM Gabby Beckford's audience. I'm a 24-year-old female engineer, often the only woman on my team/company. Paid least, proved most. Haven't taken a vacation in 2 years. I save every travel post, screenshot every tip, and do absolutely nothing about it.

That gap between inspiration and action? That's where Delusional Leap lives.

Gabby's Brief #3 asked for an app that helps her audience stop scrolling and start doing. I didn't just build it I built it because I needed it.

Before writing a single line of code, I spent a week watching every Gabby Beckford video I could find, Ted talks, YouTube vlogs, Instagram reels. I read her blog posts, scraped her community engagement, analyzed who her audience actually is, their age, their mindset, their fears. These are women who dream big but freeze at the starting line. They don't need another to-do list. They need something that feels like a friend nudging them forward.

Then I spent another week designing every screen to feel warm, approachable, and empowering — not like a productivity app, but like a conversation with someone who believes in you. Every word, every color, every interaction was crafted so a woman opens it and thinks "this gets me."

But building it also meant asking women to share personal goals, dreams, and proof of action. That's a lot of trust. So security wasn't an afterthought it was the foundation. AI verification is designed so your data is never sent raw to any model. It's encrypted, stripped of identifiers, and processed securely. Because Gabby wants you to be delusional, not defenseless.

What it does

For Users (Gabby's audience):

  • 3 daily Moves: Quick (+1pt), Power (+3pt), Boss (+10pt) — swipe-based card UI like a dating app for your dreams
  • AI-powered proof verification — Boss Moves require real proof, verified by AI via OpenRouter. No faking it.
  • Confetti celebrations on every single completion
  • Win Cards — shareable milestone cards branded with @delusionalleap + @packslight for organic viral growth
  • Streak tracking, vision board, milestone stamps (Explorer, Trailblazer, Pathfinder, Main Character)

For Gabby (Creator Dashboard):

  • Live at delusional-leap.vercel.app
  • Add/edit daily tasks by tier — she controls what her audience works on
  • Create sponsored brand challenges
  • Real-time analytics — completion rates, user growth, engagement metrics
  • Export brand reports for sponsorship deals
  • Notion-style UI with real-time Supabase sync, tasks appear on users' phones instantly

RevenueCat Integration (12 Features): SDK, Entitlements, Offerings, 3 Subscription Tiers ($4.99/mo, $39.99/yr, $79.99 lifetime), Custom Paywall, 7-Day Free Trial, Strategic Placement (paywall triggers at peak dopamine after 3rd Move), Customer Center, Restore Purchases, Streak Freeze Consumable ($0.99), Premium Gating, Real-Time Customer Info Listener.

How we built it

  • Mobile App: React Native + Expo, TypeScript, Zustand state management, expo-image-picker for proof photos, react-native-confetti-cannon, react-native-view-shot for Win Cards
  • Backend: Supabase (PostgreSQL + Realtime subscriptions + Row Level Security)
  • Creator Dashboard: Next.js 14, Tailwind CSS, deployed on Vercel with auto-deploy on git push
  • AI Verification: OpenRouter API with Llama 3.1 8B — proof text is stripped of PII before processing
  • Monetization: RevenueCat SDK with full entitlement and offering management
  • Design: Hibiscus coral (#FF3366), deep plum (#1B0A2E), cream (#FFFBF5), sunset gradients. Fraunces for headlines, Sora for body. Every screen designed to feel like talking to a friend, not using an app.

Challenges we ran into

  • RevenueCat test API keys crash real APK builds — had to configure a Google Play Store app in RevenueCat for production keys, which wasn't obvious from the docs
  • Supabase real-time sync between the creator dashboard and mobile app required careful RLS policy design so Gabby's task updates appear instantly on users' phones
  • AI verification needed graceful offline fallback — when API calls fail, users can still complete moves without being blocked
  • Building both a consumer mobile app AND a full creator dashboard with analytics in hackathon timeframe was the biggest challenge, two complete products shipping simultaneously.

Accomplishments that we're proud of

  • Two complete products — a polished mobile app AND a full creator dashboard, both live and connected in real-time
  • AI proof verification — no other goal app on the market verifies that you actually did the thing. This is a first.
  • 12 RevenueCat features deeply integrated — not just a paywall slapped on, but strategic placement at the moment of peak user delight
  • The research — spending a full week understanding Gabby's audience before writing code meant every design decision was intentional, not guesswork
  • Security-first design — data stripped before AI processing, with a clear roadmap to TEE environments for zero-knowledge privacy at scale
  • It works on a real phone — APK built, tested, deployed. Not a prototype. A real app.

What we learned

Smart women don't need more inspiration. They need accountability with proof. And creators like Gabby don't just need an app, they need a business platform with analytics, sponsorship tools, and audience insights.

I learned that asking women to share their dreams comes with the responsibility of protecting them. That's why AI verification strips identifiers before processing, and why TEE environments are the future real-time computation where not even the developer can access raw data.

On the technical side his was my first time building with Expo and EAS Build. Going from basic Expo experience to a production APK on Google Play Internal Testing was a steep but rewarding learning curve. The managed workflow is powerful once you understand its constraints.

What's next for Delusional Leap

  • H100 GPUs running open-source AI for verification at scale, no API dependency
  • TEE (Trusted Execution Environments) so women's data stays private even during computation with zero-knowledge proof verification
  • Brand analytics without personal data — sponsors get engagement metrics without ever seeing individual user information
  • Gabby onboarding — if she takes this forward, a guided setup where she configures her brand, tone, and first 30 days of Moves and scaling the application to handle about a 1M users.

Built With

Share this project:

Updates