The Fairplay system allows any distributed protocol to be specified as if a trusted external party exists. It implements the specified protocol securely as if the trusted party existed, but without relying on any such party in reality but rather on provably secure cryptographic "secure function evaluation" protocols. The system includes a high-level language for specifying a distributed protocol, a compiler that compiles the high-level definition into a low-level sequence of primitive operations, and cryptographic protocols for securely executing this sequence of operations.

The first version of Fairplay was intended for two-party computation and is described in a paper published in USENIX security symposium 2004.

The second version, FairplayMP was intended for multi-party computation and is described in a paper published in ACM CCS 2008.