This project brings to life a fake news classifier, by deploying it using Flask on Heroku. It was also successfully deployed on Azure Portal with no adjustments. The classification model is inspired by work from Aaron Edell, whose article inspired me for this project.

The article title is: "I trained fake news detection AI with >95% accuracy, and almost went crazy"

The WebApp can be found following this link. It is deployed using Flask on Heroku, and can be deployed on Azure Portal with no further adjustments. The Web App returns a probability a given article is fake news or not.

NB: The app might not be available at a given time, as Heroku turns them on and off to preserve resources.

Layout is built using HTML & CSS.

1. A text box collects the user article:

2. Returned prediction:

The user can then decide to test another article.