This document summarizes a technical seminar about Grouper, a system that provides dynamic document clustering of web search results. Grouper clusters search results using the STC algorithm to group similar documents together and make the results easier to browse. An evaluation found that Grouper produced more coherent clusters than k-means clustering and that users viewed more documents, spent less time per document, and had shorter distances between clicks when using Grouper compared to a standard ranked search results list. The conclusion discusses some limitations of the initial Grouper system and proposes improvements in a next version.