SavorSight

SavorSight turns a fridge photo into a personalized recipe—helping users cook smarter and waste less.

Inspiration

We were inspired by a common everyday problem: people often have food at home but do not know what to cook with it. SavorSight helps turn overlooked fridge ingredients into a useful meal, reducing decision fatigue and food waste.

What it does

SavorSight lets users upload a fridge photo, choose dietary needs, cooking time, and kitchen confidence, then receive a personalized recipe. It identifies visible ingredients, suggests swaps, and includes a food-safety reminder.

How we built it

We built SavorSight as a responsive web app with HTML, CSS, JavaScript, and a lightweight Node.js server. GPT-5.6 vision receives the fridge image and preferences through the OpenAI Responses API, returning structured recipe data that powers the interface. We also added demo mode so the project can be tested without an API key.

Challenges we ran into

The biggest challenge was making AI output consistent enough for a clean product experience. We solved this by requesting strict JSON from the model. We also needed to avoid unsafe assumptions, so the prompt asks the model to identify only reasonably visible ingredients and remind users to verify them.

Accomplishments that we're proud of

We are proud of turning an everyday problem into a focused, polished prototype. SavorSight combines a visual upload flow, personalized preferences, recipe generation, ingredient confidence labels, substitutions, and demo mode in one simple experience.

What we learned

We learned that AI is most helpful when it supports a specific decision rather than generating generic content. We also learned the importance of structured outputs for dependable AI interfaces. Codex accelerated the build process, while we directed the product scope, design, and user experience.

What's next for SavorSight

Next, we want to add expiry-date awareness, pantry tracking, saved recipes, nutrition estimates, and multiple recipe options. We also want users to be able to scan grocery receipts and receive proactive suggestions before ingredients go to waste.