Techniques for selection of content within web documents are described herein. In implementations, a selection tool for a browser or other client application is configured to analyze web documents to recognize different regions in the arrangement of content for the document. The selection tool may parse the underlying structure of a webpage (e.g., the mark-up code and DOM structure) to identify different discrete regions in the webpage. In addition or alternatively, the selection tool may use optical recognition techniques to visually analyze the webpage as it appears when rendered and detect different discrete regions based on the visual analysis. Recognized regions are configured as portions of the content that are individually selectable. These selectable regions may be exposed and made accessible in various ways to enable users to toggle selection of the regions quickly and easily.