Inspiration

Have you ever stayed in an Airbnb and opened countless cupboards to look for a cup, or countless drawers to look for a pan? That's been my experience, and at one point I was tempted to text the host late at night to ask. But I didn't want to disturb them just to find a frypan. But what asking was easy?

What it does

OpenDrawer indexes all the rooms and items of a house or apartment. Hosts simply take pictures of their property, optionally add text information like WiFi passwords, and they get a context-aware assistant for their guests. Guests can ask anything about the property. Examples:

  • Where is the coffee maker?
  • Where are the cups?
  • Is there anything to read, like magazines? (The bot might see magazines somewhere in the house)
  • Where is the kayak?
  • Where do I take out the garbage to?
  • What is the WiFi password?

How I built it

  • Vercel for hosting
  • Gemini API for chatbot requests
  • Supabase
  • Cursor

Challenges we ran into

  • Originally, my plan was to make a pet collar with a camera. We brought hardware but it didn't work. My teammate wanted to continue with a phone as the camera, while I wanted to pivot to this to gain experience deploying something. We both decided to switch to the solo stream, which was best for us to both get what we wanted out of the hackathon.
  • I had to refine the AI bot prompt a few times.

Accomplishments that I'm proud of

Deploying a fully usable app, with quick setup for hosts and seamless use for guests.

What I learned

How to use Cursor - I love it! Learned Supabase and Vercel deployment. Also learned about UI/UX design as that was a big focus of this project.

What's next for OpenDrawer

AirBnB integration so hosts can directly pull your photos from their listings.

Built With

Share this project:

Updates