For companies active in the domain of developing data-driven and proprietary algorithms, a big challenge is always how to share or sell their solution to other companies without exposing the internal logic. This problem is more highlighted, especially in the use cases where the data processing should be done on the edge and in real time, even in remote areas, eliminating the possibility of using a server-side logic. The solution presented in this work provides a reliable way to securely share the physiological prediction algorithms for third-party mobile applications, in the domain of sleep, fatigue, and Alcohol misuse. It consists of a Swift XCFramework for iOS and an Android AAR library, each providing a clear and consistent public interface while keeping the logic secure and unreachable.