Smart devices are attractive platforms for researchers to collect data coming from several sensors due to their small size, low cost, and the fact that they are already carried routinely by most people. The capability of smart devices to be used as the target of a positioning system has been already demonstrated in previous works. However, most of them rely on a single technology, or they are specific to the environment or user. In this paper we tackle these constraints by presenting a novel seamless positioning system which fuses the sensors information provided by a portable smart device to perform real time location without interruption and independently of the environment the user is moving. We have tested the system with a commercial smart device in an uncalibrated three floor building and its surroundings fusing the GNSS, WiFi and barometer as frequently used sensors, and the microphone and the proximity contactless technologies as occasionally used sensors. The obtained positioning accuracy mainly depends on the indoor path-loss awareness and on the markers density, showing that without using markers but dynamically estimating the path-loss exponents we obtain an error of <2 m for 90 % of cases.