One of the common type of summaries are year-end retrospectives, for example those that discuss the films of a year. This paper addresses a framework that facilitates semi-automated authoring of such a summary for the domain of film based on material gathered in web-based knowledge sources like the Internet Movie Database (IMDb). The aim is twofold. First the system should facilitate a journalist with an audio-visual summary on which a finalized review show can be established. Second, the system should also be made available in a web environment, such as the IMDb, so that the general public can generate trailer-based shows for their personal taste. The framework introduces techniques that exploit different data sources, such as the IMDb, Wikipedia, and FilmVandaag for the automatic retrieval of relevant material, and template driven presentation techniques for the development of the general story outline; and the generation of the final video by applying a set of trimming rules. Example generations are presented and discussed and an outline of future work is presented.