Information is increasingly in a digital-only format and almost everything we do nowadays depends on a digital identity for authentication and authorization. No matter how strong the authentication system is, after the authentication phase, there is no continuous verification that user is still the same human being that successfully logged in, thus leaving the system unprotected. This paper presents a usable breakthrough approach for continuous authentication with free typed text through the use of biometric techniques based on keystroke dynamics. Our main purpose is to achieve a reduction of the required sample size, while improving (or at least not worsen) precision performance, by adapting and improving parameters on a keystroke dynamics algorithm.