The resource allows an access to a trained neural network on around 30000 examples of propagandistic and non-propagandistic articles written in English language.

To test an article, send a link or a text in a form above. The result is a number in the range from 0 to 1, where 0 means that the text does not resemble the propaganda at all.

For text, the neural network additionally grades each paragraph and marks the words that are the most important in the context of the prediction.

If you think the grade is eggregously wrong, please press "disagree" to help us improve accuracy of predictions.