sharables

inspiration

do you love eating out with friends, but find it hard to agree on where to go? have you ever wished you could easily find restaurants that match everyone’s tastes, or share dishes you couldn’t finish alone?

sharables is an ai-powered dining companion that makes group dining simple. it helps you and your friends pick the perfect restaurant, recommends shareable menu items, and ensures everyone have a great time together over amazing food experience.

what it does

pick a restaurant that catches your eye and give thumbs up/down to the menu items you like or dislike. your preferences are stored as an embedding, which is used to:

• match you with people who have similar dining tastes
• select the optimal restaurant for your group using nearest neighbor restaurant by similarity distance

how we built it

• windsurf to vibe code
• cohere to generate menu based on recent restaurant's reviews (yay hack!) and user preferences embedding, similarity score between users, restaurant selected for the group is the one that is the nearest neighbor to the optimal restaurant for each person in the group.

challenges we ran into

no public menu api... windsurf doesn't do what i want it to do... things break every time I add something new and one-shotting doesn't work :(

accomplishments that we're proud of

debugging all night after attending hackathon activities fulltime & pulling up solo

what we learned

use cohere api for the first time; we found out that there isn't any friendly public api for restaurants' menu

what's next for sharables

• add dietary rstrictions to user embedding
• integrate a real menu api
• allow users to pick dining time & date and match with people
• improve app performance: faster loading, continuous scrolling, better frontend
• given a group of people, provide restaurant recommendations along with suggested menu items

Built With

• cohere
• firebase
• firestore
• javascript
• windsurf