Quotation Extraction consists of identifying quotations from a text and associating them to their authors. In this work, we present a Direct and Indirect Quotation Extraction System for Portuguese. Quotation Extraction has been previously approached using different techniques and for several languages. Our proposal differs from previous work, because we build a Machine Learning model that, besides recognizing direct quotations, it also recognizes indirect ones in Portuguese. Indirect quotations are hard to be identified in a text, due to the lack of explicit delimitation. Nevertheless, they happen more often then the delimited ones and, for this reason, have an huge importance on information extraction. Due to the fact that we use a Machine Learning model based, we can easily adapt it to other languages, needing only a list of verbs of speech for a given language. Few were the previously proposed systems that tackled the task of indirect quotations and neither of them for Portuguese using a Machine Learning approach. We build a Quotation Extractor using a model for the Structured Perceptron algorithm. In order to train and evaluate the system, we build QuoTrees 1.0 corpus. We annotate it to tackle the indirect quotation problem. The Structured Perceptron based on weight interval scheduling obtains an F1 score of 66 percent for QuoTrees 1.0 corpus.