Homesty.ai

A buyer asks about a property and gets answers from the real listing.

  • Next.js
  • React
  • Node.js
  • PostgreSQL
  • Prisma
  • Vercel
  • Sentry
Live

The problem

Buyers ask listing agents the same five questions about every flat, and agents answer them by hand all day. Homesty takes that first conversation: a buyer asks about a property and gets an answer taken from the real listing.

What I build

I work at Homesty.ai LLP as a software engineer, and I have been building it since November 2025. I build the side buyers see.

  • The landing page and the chat screen where a buyer asks about a property, in Next.js and React.
  • The REST endpoints behind them, in Node.js.
  • The database schema, in PostgreSQL, queried with Prisma.
  • The admin pages the team uses to manage listings.
  • The login flow, and an audit log that records what changed and who changed it.

I release on Vercel, watch production errors in Sentry, and work in Git with pull requests and code review.

What it does not do

  • It answers from listing data, so a thin or stale listing gives it little to work with.
  • It handles the first conversation. Visits, negotiation and the sale stay with people.

Links