Exploiting facts that are published online using semi-structured or unstructured formats is a highly complex task, pieces of data are typically published in isolation and periodically updated in a bulk fashion without any coordination. In this paper, we propose a new software pipeline tackling this issue and semantically lifting online facts as Linked Data in a continuous manner. Our solution, LinkedPolitics, extracts, converts, interlinks, and finally makes available as Linked Data online facts that could not be transparently exposed and exploited otherwise. Our system is online, in the sense that it automatically makes available new facts as they are uploaded online. We also present a deployment of LinkedPolitics centered around the Swiss Parliament. Our deployment extracts and semantically lifts a flurry of political facts that are periodically published in unstructured form by the Library Service of the Swiss Federal Assembly. It then automatically materializes a number of views to support different visualizations of the data in order to capture the actions, announcements or votes made by the politicians, as well as their relationships to each other and to external entities such as companies.