🍳 HostelChef
Turn whatever's in your room into something you can eat.
What I Built
I built HostelChef for a friend who often has a random collection of ingredients in their hostel room but struggles to figure out what they can actually make with them.
Instead of searching through recipes and trying to find one that matches what they have, HostelChef lets them enter:
- Available ingredients
- Cooking time
- Equipment
- Dietary preference
It then generates a simple, practical recipe tailored to their situation.
The goal was simple: make deciding what to cook in a hostel a little less painful.
Demo 🚀
Live Demo: https://hostelchef.onrender.com
Code 💻
GitHub: tushar-313/HostelChef
How I Built It
HostelChef is built around Gemma 4, Google's open-weight AI model.
The application uses:
React + Vite → FastAPI → Gemma 4
The frontend collects the user's ingredients and constraints, while the FastAPI backend sends them to Gemma with a structured prompt.
Gemma generates a recipe containing:
- Recipe name
- Description
- Required ingredients
- Cooking time
- Difficulty
- Step-by-step instructions
- A practical hostel cooking tip
I kept the project intentionally small. The focus wasn't on building another huge recipe platform, but on using an open model to solve one very specific everyday problem.
Why Does Open Innovation Matter?
For a small personal tool like HostelChef, open models make experimentation much more accessible.
Instead of building the application around a proprietary AI service and treating the model as a black box, using an open-weight Gemma model gives developers more flexibility to experiment with the underlying model and build applications around it. It also makes projects like this easier to adapt for different environments, constraints, and use cases.
For me, that's the interesting part of open innovation: a developer can take an openly available model and turn it into something useful for a very specific problem.
Prize Categories 🏆
Best Use of Gemma
Top comments (0)