Personlized Recipe Suggestor

Inspiration

Due to Wastage of Food Items due to expiry of food items , lack of prioritizing the food items according to expiry date

What it does

it takes a Image of Food items as input and recongnize the food items and do not include non edible food items in it and add any item extra , it prioritizes the food expiry and try to cook the food which will expire soon , different cuisines , exclude allergies , detailed instructions , youtube video , time , calories to user

How we built it

Gemini , Langchain , langgraph , youtube API , Vite ,

Challenges we ran into

Handing the Image , structing the output

Accomplishments that we're proud of

completed it end to end

What we learned

How to build a multi-modal AI pipeline that takes image input and converts it into structured, usable data. Handling LLM output structuring using schema enforcement and controlled prompting. Integrating external APIs (YouTube API) to enrich LLM-generated results with relevant cooking videos. Managing edge cases like: Non-edible item detection Duplicate ingredients Allergy filtering Expiry-based prioritization logic