We introduce a constrained mechanism design setting called internal implementation, in which the mechanism designer is explicitly modeled as a player in the game of interest. This distinguished player has the opportunity to modify the game before play. Specifically, the player is able to make reliable binding commitments of outcome-specific monetary transfers to the other players in the game. We characterize the power of internal implementation for certain interesting classes of games, and show that the impact of internal implementation on the utility of the players' and the social welfare is often counterintuitive; for example, the social welfare can be arbitrarily worse after an internal implementation. Categories and Subject Descriptors I.2.11 [Artificial Intelligence]: Multiagent Systems General Terms Economics, Theory Keywords Game Theory, Constrained Mechanism Design, Implementation