Inspiration
We were inspired by chaotic cooking TV shows like Hell’s Kitchen and The Bear, where success depends on timing, delegation, and coordination under pressure. We wanted to take that chaos and turn it into something structured and manageable using software.
What it does
Restaurant Commander takes a recipe and automatically breaks it into structured, ordered steps. It models dependencies between steps, assigns tasks to chefs, and tracks what each chef is currently working on, what’s next, and overall progress.
How we built it
We built the backend using Python and FastAPI, with NetworkX to represent recipes as directed graphs of dependent steps. We used AI to parse recipe descriptions into structured steps and inputs, then transformed those into a step graph that can be split among multiple chefs. The system exposes API endpoints to progress steps, assign work, and track kitchen state.
Challenges we ran into
We faced significant trouble in discovering how to connect the Gemini API to take in JSON recipes and turn the recipe into a graph for our program to order the steps for the chefs.
Accomplishments that we're proud of
Successfully using the Gemini AI model to transform plain text into graphs that our code can follow.
What we learned
We have updated our abilities in integrating AI models in VS Code using Python.
What's next for Restaurant Commander
Hardware integration, as well as designing a more fluid or easy-to-use UI.
Log in or sign up for Devpost to join the conversation.