Indoor localization is becoming increasingly important for mobile applications. WLAN fingerprinting is a compelling technique because it builds upon existing infrastructure and client hardware available in off-the-shelf mobile devices. We evaluate different methods for WLAN fingerprint classification with a focus on on-device localization. The main scientific contribution of this approach is that any Android based device can localize itself (without any server being able to determine the current location) using existing WLAN infrastructure (no additional access points have to be installed, the firmware of existing access points doesn’t have to be changed). This approach was chosen to make indoor localization feasible in non-academic use cases. With a functional implementation and a simple procedure for collecting WLAN fingerprints, we currently achieve an accuracy of 4,m in 90% of all cases with a mean error of only 2.2,m when the same device is used for training and testing. Next steps are calibration between different mobile devices, post-processing in terms of movement, and automatic downloading of the required WLAN fingerprint databases on a global scale.