In 2012, Brazilian President Dilma Roussef installed the Brazilian Truth Commission (CNV) to address gross human rights violations that occurred from 1946-1988.One of the most important sources of information available regarding this period is the files of the agencies that comprised the Brazilian intelligence system during the dictatorship. In total, there were around 12 million pages of relevant text in the National Archives. To make effective use of this trove of information, the CNV was challenged to use some data science tools to look for useful information within this huge dataset. As a result, a prototype of a data repository with selected documents (pdfs, images, etc.) has been created, which we summarize in this note. Computational tools for searching, organizing, and visualizing potentially important documents were developed and utilized to support CNV researchers. We also reflect upon the issues that complicated the CNV’s ability to gain access to reliable and comprehensive data and the limitations of analysis conducted with this type of research.