In this paper we propose the practical and secure solution how to integrate the fingerprint biometric into cryptographic applications. A binary string which could be reproduced unambiguously, practically, a cryptographic key, can reliably protected with originally fingerprint codes. The key is generated using random number generator, such as HSM device. In that way we are keeping coincidence of cryptographic key, also keeping its entropy, and what is more important if the key is compromised, we can easily use the new one. For the key encoding we are proposing Slepian Wolf coding, basically defined in Slepian Wolf theorem. Like a method for converting fingerprint into minutiae and after that to a binary string we propose Rajarshi Guha method, detail encoded in C programming language and approved in practice. After combining two mentioned binary strings we get an unique data set and place them onto the chip card. The some other kinds of tamper resistant tokens can be used as well. The key revelation process is strictly conditioned by the presence of fingerprint and the chip card. Proposed solution provides maximum security of fingerprint and the lowest possible errors that can be occurred. Also the degree of false positive and false negative authentication is reduced to a minimum.