Abstract: Methods, systems, and computer-readable media for recommending completions for partial strings while a user is typing in characters of the partial strings are provided. A search engine executed by a server receives the partial string. In response, the search engine generates autosuggestions and user-specific completions. The user-specific completions are based on a history corresponding to the user. The user-specific completions are also associated with one or more URIs. The autosuggestions and user-specific completions are transmitted in display rank order to the user. The autosuggestions and user-specific completions are displayed in a dropdown box that allows the user to select a completion for the partial string. A URI corresponding to the user-specific completions is displayed based on user interaction with the user-specific completions include in the dropdown box.

Abstract: Filters are generated by analyzing a set of search results most relevant to a search query to determine filter values that may be of most interest to the user. A search query is received and is classified into a query segment that best corresponds to the search query so that filter categories associated with the query segment can be identified. Based on an analysis of search results most relevant to the search query, metadata associated with these search results is used to formulate filter values for each filter category. The filter values are ordered based on their respective frequency of presence in the search results. The search results, along with the filter categories and filter values, are communicated for presentation.