Inspiration
Eating food that is affordable, healthy, and enjoyable can be challenging. As students, it is especially hard to find meals that keep both you and your wallet healthy. RecipeFind was inspired by this problem. It helps users select nearby grocery stores and discover recipes that match their dietary restrictions while using ingredients that are currently on sale.
What it does
RecipeFind parses weekly flyers from many Canadian grocery stores to find deals on a wide variety of ingredients. It then leverages AI to generate recipes that use these discounted items while respecting the user’s dietary restrictions.
How we built it
RecipeFind uses Yellowcake to parse multiple grocery store websites and extract weekly deals. This data is exposed as a tool to Solace’s Agent Mesh, which coordinates several agents responsible for generating recipes and filtering them based on dietary needs.
Challenges we ran into
Using agents throughout the entire stack, from website parsing to recipe generation, introduced significant latency. We were able to reduce processing time by about 50 percent by caching parsed flyer data in a Redis database, which greatly improved the speed of finding items on sale.
Accomplishments that we’re proud of
We built a very usable application that could genuinely help people like ourselves make better food choices on a budget.
What we learned
During the hackathon, we became more familiar with new offerings from Solace and Yellowcake. We also gained additional experience working with Docker Compose, including running both the Solace Agent Mesh and a Redis container together.
Log in or sign up for Devpost to join the conversation.