The present description discloses a technique for recovering data using a timeline-based computing environment. Data items of the application are periodically saved for recovery such that the saved data items can be used to recover the application at a point in time when the items are saved. As a result, a search through a time-based computing environment is provided to recover the application at different points in time. The application with the saved data items can then be recovered at a designated point in time. Each saved data item can also be indexed with metadata, which are used to conduct a search to generate a list of data items according to a match between the indexed metadata and a user selected variable. Moreover, when the application is a communication client having multiple messages, an index data to indicate whether a message in the communication client is spam is saved. Using this index data, a search that includes or excludes the spam messages can then be conducted. Thus, using a timeline-based computing environment, an improved method and system of recovering data that is more user friendly, effective, and manageable is thus provided.