</>
{ }
npm run build
[ compiling... ]
Developer Workspace
Fetching Projects...
Loading Components...
Building Experience...
Preparing something amazing for you.
compiling100% ready soon
Full Stack AI PlatformFeatured Project
NextMatch
AI-Powered Property Rental Platform
A full-stack rental discovery platform where users browse, filter, and list properties, enhanced by two agentic AI features: a context-aware chat assistant and a feedback-driven recommendation engine.
Enlarge Screenshot
Project Screenshots (6)
Cover
Full Project Description
NextMatch reimagines property rental discovery by combining a polished full-stack marketplace with genuine AI reasoning. Users can search and filter verified listings, publish their own properties, and manage them through a personal dashboard — all secured with Better Auth and Google OAuth. What sets it apart is the AI layer: a conversational assistant that understands live listing data and answers grounded, context-aware questions with full conversation memory, alongside a smart recommendation engine that ranks properties against user preferences and continuously improves from thumbs up/down feedback.
Technologies Stack
Next.jsTypeScriptExpress.jsMongoDBBetter AuthGemini API
Key Features & Functional Highlights
- AI Chat Assistant with persistent conversation history, live database-grounded answers, and suggested follow-up prompts
- Smart Recommendation Engine that ranks listings by user preferences and improves from ongoing feedback
- Full listing CRUD with search, multi-field filtering, sorting, and pagination
- Secure authentication via Better Auth with Google OAuth and a one-click demo login
- Fully responsive dark/light theme UI across homepage, explore, details, and dashboard pages
Technical Challenges Faced
- Resolving cross-domain session cookie handling between a separately deployed Next.js frontend and Express backend, including SameSite, Secure, and __Secure- prefixed cookie naming in production.
- Migrating from a deprecated Gemini SDK and model version mid-development while preserving streaming responses and conversation memory.
Future Roadmaps & Planned Improvements
- Adding a map-based explore view using existing listing geolocation data.
- Expanding the recommendation engine with a persisted user favorites/saved listings feature.