This post is my submission for DEV Education Track: Build Apps with Google AI Studio.
What I Built
AI is evolving at a very fast pace. Now, it’s possible to bring ideas to life quickly thanks to the power of Google AI Studio, using Gemini as the core engine that turns prompts into working code. With well-structured, concise prompts, we can generate almost any application to solve problems or provide intelligent suggestions.
For my project, I created a responsive web application that acts as a virtual supermarket shopping assistant. The idea is to allow users to take a photo of their fridge or pantry, detect what's inside, identify missing items, and even suggest recipes based on available ingredients. It also supports manual entry of products and includes multilingual support (English and Spanish) via a constant defined in TypeScript.
Prompt Used
Create a virtual supermarket shopping assistant that works with computer vision. This assistant should allow the user to take a picture of their fridge or pantry and, from that image:
- Identify the visible products (food, beverages, condiments, etc.).
- Detect products that are close to expiring or look spoiled (such as overripe fruits or vegetables).
- Determine which products are in short supply by comparing with a frequent or suggested shopping list.
- Automatically generate a shopping list with missing or expired products.
- Suggest quick recipes based on what is available.
The assistant must:
- Recognize common brands and packaging.
- Estimate the remaining quantity if possible (e.g., half a liter of milk).
- Learn user preferences over time.
- Offer integration with online supermarket platforms or export the list as text.
- Support multilingual functionality (English and Spanish).
- Allow manual addition of products.
Demo
👉 You can view the working prompt on Google AI Studio here:
🔗 Try the demo
My Experience
The experience with Gemini in Google AI Studio has been smooth, practical, and fast. During the process, Gemini was able to:
- Correct and refactor its own code,
- Validate logic and functionality,
- Perform internal testing,
- And generate deployable solutions.
Projects can be deployed directly to Google Cloud or downloaded to run locally. The code is transparent and easy to understand, which makes this a great tool for rapid prototyping and experimentation.
💡 Happy coding and building with AI!