Summar is an automatic summarization system developed for the Polish language. The system implements trainable sentence-based extractive summarization technique, which consists in determining most important sentences in document due to their computed salience. Neural networks were used as a machine learning algorithm. This is the first solution using Polish Summaries Corpus for training, therefore it can serve as a baseline for systems developed in future. The system was implemented by Łukasz Pawluczuk during work on Master Thesis's in computer science at Faculty of Mathematics and Computer Science, Adam Mickiewicz University.