This document describes a proposed food recommendation system called Sehat Co. that uses image processing and deep learning techniques to predict ingredients from uploaded food images. It then matches the predicted ingredients against a database of ingredients known to fight COVID-19 to display relevant recipes, stats on COVID-fighting ingredients, and a recommendation for a second recipe. The system aims to help users track their food intake and choose meals containing ingredients that may increase immunity against COVID-19. It was developed using the Recipe1M dataset containing over 1 million recipes and trained on a multi-task deep learning model to perform ingredient prediction and recipe generation and recommendation.