• A vectorial representation for every ingredient and recipe was generated using Word2Vec. An SVC model was trained to return recipes’ cuisines from their set of ingredients. South Asian, East Asian and North American cuisines were predicted with more than 73% accuracy. African, Southern European and Middle East cuisines contain the highest number of cancer-beating molecules. Finally, it was developed a web application able to predict the ingredients from an image, suggest new combinations and retrieve the cuisine the recipe belongs, along with a score for the expected number of negative interactions with antineoplastic drugs (github.com/warcraft12321/HyperFoods).