The application created for the bachelor's thesis, which uses face-api.js to recognize human emotions. The application implements video analysis from a webcam source and uses a peer to peer network to connect two clients to each other to analyze each other's emotional reactions.

The application works in two web browser windows at the same time and has a presenter panel and an audience panel. The videos from the two cameras are analyzed at the same time. Messages appear in the live presenter's panel suggesting what the presenter can improve in his speech. The messages are selected based on the analyzed emotions among the audience and the behavior of the presenter.