HomePilot

An AI rental assistant for first-time renters: it monitors listings, predicts your approval odds, spots scams, and walks you through applications so you don’t have to figure it out alone.

Comment

Inspiration

Our teammate Ali was apartment hunting in Newport for the first time — no rental history, no idea how to compare listings, spot scams, or improve his chances of getting approved. He ended up doing everything manually: a sprawling Excel spreadsheet, hours of research, and a lot of anxiety. He wasn't alone in that experience. First-time renters are handed one of the most stressful financial decisions of their lives with almost no guidance.

HomePilot is our answer to that. We wanted to build the co-pilot Ali never had — an AI-powered rental assistant that doesn't wait for you to figure things out, but actively guides you through every step.

What it does

HomePilot helps first-time renters find apartments and apply with confidence. At its core:

  • Renter Score — a single number that tells you where you stand and exactly what to improve
  • Renter Passport — a guided document setup flow that bundles everything a landlord needs into a one-click application packet
  • Match Percentage — every listing is scored against your profile so you know which ones you're actually likely to get
  • Proactive AI Alerts — the app monitors listings in real time and tells you when to act, flags scams before you waste time, and surfaces urgency signals like high competition
  • Auto-Optimization — the AI suggests profile improvements ranked by impact, so you're never left guessing what to do next

How we built it

We built HomePilot as a React 18 single-page app with TypeScript and Vite, styled with Tailwind CSS v4 and shadcn/ui for a consistent, modern design system. Supabase handles authentication, our PostgreSQL database, file storage, and realtime updates — which let us move fast without building a custom backend from scratch. We used Motion for animations and Lucide React for icons throughout.

The product logic lives in custom hooks and contexts for auth and data fetching, with Row-Level Security enforced at the database level so users can only ever access their own data. We kept the stack lean and deliberate — every choice was made to let us iterate quickly on the renter experience rather than on infrastructure.

Challenges we ran into

The hardest problem was the Renter Score. We had to decide what factors to include (income, documents, credit range, rental history, references), how to weight each one, and how to surface a single number in a way that felt fair and actionable — not opaque or discouraging — especially for users who, by definition, have thin histories.

Getting the balance right between accuracy and simplicity took several iterations. A score that's technically correct but demoralizing isn't useful. We kept asking: does this help the user know what to do next? That question forced us to rethink the score not as a judgement but as a roadmap.

Accomplishments that we're proud of

We shipped a coherent end-to-end flow for a genuinely underserved user — from signup and onboarding through a Renter Score, Passport setup, listing discovery, match percentages, and proactive AI suggestions, all in one cohesive app. We're also proud of the details: working profile management, email and password changes, a theme toggle, and a visual system that holds together across every screen.

But honestly, the thing we're most proud of is that Ali's spreadsheet is now a single place. That felt like the real goal.

What we learned

We learned that first-time renters aren't just uninformed — they're overwhelmed, and a little structure goes a long way. Scores, odds, checklists, and pre-built next steps dramatically reduce the cognitive load of what is otherwise an opaque process.

On the technical side, we went deep on Supabase — RLS policies, database triggers, Edge Functions, and Realtime subscriptions — and on building a design system that scales cleanly with Tailwind and shadcn. We also learned the value of tight scoping: picking one clear persona and a few high-impact features, and building those really well, beats trying to do everything at once.

What's next for HomePilot

  • Smarter Renter Score — more signals, clearer per-factor explanations, and market-specific calibration
  • Deeper listing coverage — more sources, better filters (pet-friendly, student-friendly, short-term), and verified landlord data
  • Application tracking — reminders, status updates, and a timeline so users never lose track of where they stand
  • Livability and neighbourhood insights — crime index, rent inflation trends, transit scores, and demand heat per property
  • More personalized alerts — the right nudge, at the right moment, based on what each user actually needs

Long term, we want HomePilot to be the single trusted co-pilot for anyone renting for the first time — not just a tool you use once, but something that grows with you through every move.

Built With

Share this project:

Updates