AI BNB reimagines Airbnb as if it were founded in 2025 — powered entirely by AI.
A chat-first platform where travelers describe their ideal trip, and AI plans, finds, books the perfect stay, and hosts can generate listings from just a few photos.
- 🎥 Loom Demo Video →
- 🧠 Built at Hack Princeton 2025
As travelers, our group loathes the planning process that comes with long trips — from endlessly browsing Airbnbs to manually piecing together an itinerary.
We also sympathize with home renters, who spend hours uploading images, writing descriptions, and managing bookings. So we built AI BNB — an all-in-one AI utility that makes life easier for both travelers and hosts.
- Chat with an AI agent about destinations, group size, budget, and activities.
- Receive personalized property and itinerary suggestions.
- Collaborate with friends to co-plan and vote on trip ideas.
- Upload photos, and the AI generates the full listing (title, description, amenities, price).
- Preview and publish instantly
| Layer | Technology |
|---|---|
| Frontend | React + Next.js + Tailwind CSS |
| Backend | Flask (Python) + Dedalus Labs Agents |
| Database & Auth | Supabase (PostgreSQL + Auth + Storage) |
| AI & Tools | OpenAI GPT-4o + Dedalus MCP servers |
| Containerization | Docker |
Frontend (Next.js + Tailwind)
│
├── /api/query-agent → Flask Backend
│
Backend (Flask + Dedalus)
├── Tools:
│ ├── extract_trip_info
│ ├── clarify_missing_info
│ ├── find_stays
│ ├── book_stay
│ └── create_listing
│
Database (Supabase)
├── users
├── listings
├── bookings
└── experiences
- Learning and implementing agentic AI workflows
- Integrating Flask, Supabase, and Next.js seamlessly
- Balancing design, backend logic, and AI orchestration between two people
- Built a fully functional AI travel agent that books realistic stays
- Designed a host-side AI generator that turns raw photos into listings
- Created a chat-first UX that feels human and futuristic
- Learned and implemented Dedalus Labs tools for orchestration
- Delivered it all in under 36 hours 🚀
- How to design for AI-first UX instead of traditional search
- How to manage multi-step reasoning with agentic AI
- That collaboration and asking for help are key to big projects
- That AI can simplify one of the most painful workflows — trip planning
- 🧭 Live itinerary updates based on weather or flight changes
- 🧩 More integrations (Booking.com, TripAdvisor, Google Places)
- 💳 Stripe payments for real bookings
- 📱 Mobile app for iOS and Android
Our mission is simple: Turn the pain of planning into the joy of anticipation.
# Clone repo git clone https://github.com/<yoge1212>/ai-bnb.git cd ai-bnb#set env var
#docker-compose up
| Name |
|---|
| Yogesh Sampathkumar |
| Jessie Singh |
Special thanks to Dedalus Labs, and our mentors.


