Nisepu

Real Estate AI

Comment

🏠 Nisepu: Smart Real Estate AI App for Nigerians 🇳🇬

🚀 What Inspired Me

As a Nigerian developer, I’ve seen firsthand how difficult it is for people to find reliable and verified real estate listings. Scams, outdated data, and a lack of access to trustworthy agents plague the industry. I wanted to solve a real problem that affects millions — and that’s how Nisepu was born. My mission was to create a mobile app powered by AI that connects Nigerians with verified properties in a fast, smart, and transparent way.

🧠 What I Learned

While building Nisepu, I’ve learned a ton about:

  • AI integration using Google’s Gemini for smart property search and chat-based assistance.
  • Mobile development using Expo (React Native) to make cross-platform deployment easier.
  • Backend infrastructure with Supabase, which allowed for seamless user auth, real-time data, and storage.
  • User-centric design by studying common pain points in the Nigerian real estate journey.
  • The importance of scalability and performance for mobile-first users in Africa.

🛠️ How I Built It

Here’s a breakdown of the tech stack and workflow:

  • Frontend: Developed using React Native with Expo, providing a smooth cross-platform experience. The UI is clean, modern, and designed to be easy even for first-time smartphone users.

  • Backend: Powered by Supabase, I’m using:

    • Supabase Auth for user registration/login (email and phone number)
    • Supabase Database (PostgreSQL) for storing property listings, user profiles, messages
    • Supabase Storage for hosting images and documents (e.g., land titles)

  • AI/Chatbot: Integrated Gemini AI to assist users with:

    • Property search via chat ("Find me a 3-bedroom in Abuja under ₦40M")
    • Onboarding questions
    • Explaining real estate terms and legal advice in plain English or Pidgin

  • Other Tools:

    • React Navigation for routing
    • Lottie for splash animations
    • Zustand for global state management
    • Cloudinary (possibly integrating soon) for image optimization

😤 Challenges I Faced

Building Nisepu wasn’t smooth all the way. I ran into several challenges:

  • 📱 Expo limitations with native modules (like document scanning or offline maps), which I had to find workarounds for or delay until EAS Build.
  • 🔐 Auth flow bugs with Supabase’s phone OTP login, especially when testing across devices with slow networks.
  • 🤖 Prompt tuning Gemini’s AI so it wouldn’t hallucinate or suggest non-existent properties.
  • 📶 Connectivity issues: Since many users have slow/unstable networks, optimizing load time and offline support was crucial.

🎯 What’s Next (Final 20%)

I’m about 80% done, but here’s what I’m finishing up:

  • [ ] Integrating payment APIs for deposits/reservations
  • [ ] Agent verification via CAC and NIN validation
  • [ ] AI voice assistant in Hausa/Yoruba/Pidgin
  • [ ] Deploying to Play Store and getting early user feedback

🧩 Final Thoughts

Building Nisepu has been an incredibly meaningful project. It’s not just code — it’s a solution tailored for the Nigerian people. I believe this project can make real estate more accessible, transparent, and secure for everyone. I’m excited to launch and iterate based on user feedback.

Share this project:

Updates