Inspiration
Healthy eating shouldn't be a luxury. Yet millions struggle to balance nutrition with tight budgets, lacking the tools to make informed decisions at the grocery store.
What it does
Our app can plan a meal and create your shopping list the way you like. Scenarios:
- Plan a healthy meal for me with at least 50g of protein in it. Budget is around 5e: The application will combine different grocery discounts around you and will plan a shopping list as well as a detailed plan on where to buy those items.
- Plan a whole week of meal prep for me. The goal is to buy everything in one place. Find the cheapest store around me for this grocery list. The app will calculate the sum of each item in your list and compare prices in stores near you. Then it will display a detailed plan.
- Dashboard Our app analyses your shopping habits, your most searched items and creates a profile of you. Then the app gives you different meals to try with the goal of improving your health. It will also analyse your spending habits and will recommend considering going to another store next time.
How we built it
The whole app is a custom RAG with a built-in agent that has a limited amount of tools: +find the cheapest products +analyse a person's eating habits +analyse a person's spending habits +make a meal/meal prep and similar +scrape shops via connected MCP +analyse the current deals in the stores and find the cheapest/fastest/or some specific target(find bio products)
We used Qdrant as our vector database for searching for groceries We used Apify scraping MCP for Lidl, Tesco and wrote our own scraper for Fresh We used the Google maps api to identify users' locations and guide them to the nearest store.
Challenges we ran into
Difficulty in creating a UX/UI that will differentiate us from other Chatbots. Also, scraping Slovak websites proved challenging due to constant language conversion problems and hosting blocking us
Accomplishments that we're proud of
Complex RAG system with real scraped data. Also Agentic system with tailored tools for users' use cases
What we learned
We gained knowledge in MCP usage, React development and scraping. Also, we learned to use Agentic systems.
What's next for Lorem Ipsum-01
20 hours of sleep to compensate for this night:))
Built With
- apify
- google-maps
- llm
- mcp
- python
- qdrant
- rag
- react
Log in or sign up for Devpost to join the conversation.