Zeff | Devpost

Slide 3: Problem
Slide 4: Painpoints
Slide 2: Project description
Slide 1
Slide 5: Impact

Inspiration

I began to code when right before the explosion of AI in 2020. With the help of AI, I was able to learn complex concepts in an excelerated rate. I want to bring that same fuel to help people learn to cook sustainably saving them time, money, and stress.

What it does

We help the user manage their incoming ingredients, create tailored recipes, schedule meal plans and guide the user in the cook with voice feedback.

How we built it

We use the Model Context Protocol to analyze user queries, retrieve context, choose the appropriate tools, and run sandboxes to deliver rich UI components for the user to use in their cooking journey.

Challenges we ran into

The AI alone had no context to give tailored recipes and it was expensive to make multiple recipes.

Accomplishments that we're proud of

We were able to use Daytona to optimize giving context backed responses and lowering the token cost of creating multiple recipes.

What we learned

We were trying to use Daytona for tasks that didn't need to be ran in a sandbox, slowing down our response time. We were able to re-adjust and identify when a sandbox was actually needed