SecretFlow
An open-source Python framework for privacy-preserving data analysis and machine learning, with MPC, homomorphic encryption and trusted execution environments.
SecretFlow is an open-source Python framework for teams that need to analyze data or train machine learning models across separately held datasets. It brings several privacy-preserving computing methods into one framework, so applications can combine them according to their security assumptions and performance needs.
It supports secure multiparty computation (MPC), homomorphic encryption (HE) and trusted execution environments (TEE), alongside computation on plaintext data. A shared abstraction lets higher-level analysis and learning workflows work across these technologies, reducing their dependence on a particular underlying method.
Data can be split horizontally, with different parties holding different records, or vertically, with different parties holding different attributes. SecretFlow also supports hybrid arrangements. These distinctions matter when collaborating organizations hold complementary datasets rather than matching copies of the same information.
The workflow layer covers data processing, model training and hyperparameter tuning. Its computational graphs connect work across the supported computing methods, giving developers a common foundation for analysis and training instead of requiring a separate application structure for each method.
SecretFlow uses the Apache 2.0 license.
Logotyper
