SellEasy.ai - sell your old stuff with Gemini AI!

SellEasy.ai negotiates with buyers, creates product listings, crops photos, suggests prices, and assesses product quality- a Gemini-powered secondhand sales assistant!

Comment

Inspiration

Selling secondhand goods can be a pain. From creating an effective product listing to haggling over a price, there's a lot of tedious effort involved, so instead let Gemini-powered SellEasy.ai do the heavy lifting for you!

What it does

SellEasy.ai uses the power of Gemini to:

  • identify items you might want to sell in a photo
  • crop images
  • create product descriptions
  • assess quality
  • suggest prices
  • negotiate with buyers!

How we built it

Entire built using Google tools on the Google tech stack!

  • Google Antigravity for design, dev, test, and even production deployment. Barely a single line of code was touched by a human (though rigorously overseen).
  • Nano Banana was used for the logo and design flourishes.
  • Backend stack is all Google: GCP, Firebase, Firestore, Firebase App Hosting, Firebase Functions, Google Secret Manager.

Challenges we ran into

  • Deployment to prod from Anitgravity using Gemini throws up some curveballs and requires quite a lot of careful and specific prompting.

Accomplishments that we're proud of

  • SellEasy.ai is 100% live at SellEasy.ai
  • The product is built entirely using Google products on the Google tech stack
  • SellEasy.ai is functional and fun!

What I learned

  • Gemini is great at describing and cropping photos!
  • Anitigravity is an awesome IDE
  • Deployment to prod is still hard for AI if you require API and secret configuration, database setup, etc. This required some handholding with Antigravity.

What's next for SellEasy.ai

  • Integration with eBay, Facebook Marketplace, and other platforms
  • Allow Gemini agents to negotiate prices across all these platforms, comparing quotes.
  • Assess buyer credibility
  • Login/auth (only left out so as to keep it open and demoable for the purposes of this hackathon)

Built With

Share this project:

Updates