We tested whether imagining another context during encoding would offset context-dependent forgetting. All participants studied a list of words in Context A. Participants who remained in Context A during the test recalled more than participants who were tested in another context (Context B), demonstrating the standard context-dependent forgetting effect (e.g., Smith & Vela, 2001). Importantly, some participants imagined another mental context during encoding. Some of these participants imagined Context B during encoding, and when they were later tested in Context B or even in a completely new Context C, they did not show forgetting, confirming our predictions. Other participants imagined a new context during encoding simply by transforming the current context (i.e., by imagining that it was snowing in the room), and this likewise counteracted context-dependent forgetting. These data suggest a moderator of context-dependent forgetting. When the context surrounding a memory is largely mentally generated, context-dependent forgetting is eliminated.