🥘 Food Advisor
This repo contains a streamlit application powered by LangChain.
- A foods.com recipes dataset is contained in a vector database hosted in Pinecone.
- The application returns the most suitable recipes according to the user's preferences.
- The user inputs human-language text explaining what he wants for his meal.
- The LLM interprets the user's taste and elaborate a string that will be used to query the vector database.
- The similarity search yielded by the LLM query will return the most interesting recipes.
There's a snippet of the app below.
