Abstract The Belgian Agendas Project started in 2001 and has been funded by four different research grants. Initially, the Belgian Project used a codebook other than the common CAP codebook widely used right now. The bulk of the data has been coded manually and now mostly covers the 1999–2010 period (with many datasets going back to earlier periods). A peculiarity of the Belgian data is the fact that it comes from two, largely separate political systems that ‘clash’ on the national level. Party and media systems, for example, are entirely separate. The Belgian media agenda is coded particularly extensively, as the Belgian team started from an interest in the media’s effect on politics. The Belgian team has also played a pioneering role in the party manifesto coding and the protest coding.