We have a plethora of food recipes on the internet when we look for what we want to cook, which makes our jobs difficult when we only have selected ingredients at our disposal. It is easy to find a recipe and then get the ingredients to make it, but the problems occur when we only have select few ingredients, and a person wants to find what they can make out of it. In order to tackle this challenge, we create a dataset that contains a humongous set of recipes and has step-by-step guide on how to make them. A user will need to input the picture, either directly uploading them via a web interface or choosing them from the gallery after photos are uploaded. These input photos are analysed, and a recipe set gets returned where users can choose which one they want to make. It can be achieved using CNN. CNN classifies the ingredients and puts them in separate classes based on the training. The set is then matched with available database of recipe and presented in form of links and images from where user can chose what to make. Several components will interact to give recommendations and present the aesthetic UI/UX design result. Users can interact with the app without knowing all the processing involved in the background that contains data pre-processing, data cleaning, model training, and other technical things.

Keywords : CNN, Ingredients, Recipe, UI/UX, Web interface.