Precise occupancy information is extremely valuable for energy savings and optimization in the built environment. Building management systems (BMSs) can automatically turn off the HVAC systems in unoccupied spaces and adjust the ventilation rate based on the number of occupants in each zone for energy saving. Popular commercial occupancy detection and crowd counting approaches, such as the infrared sensor, camera, and mobile devices, either require the deployment of extra infrastructure or the cooperation of occupants to carry dedicated devices, which are intrusive and inconvenient for pervasive implementation. In this paper, we propose WiFree, a novel device-free occupancy detection and crowd counting scheme using only commercial WiFi enabled the Internet of Things (IoT) devices. Firstly, we design a novel IoT platform with which the fine-grand channel state information (CSI) measurements can be obtained directly from pervasive IoT devices. Given the CSI data, we propose an effective occupancy detection scheme by measuring the shape similarity between adjacent time series CSI curves. For crowd counting, we first design an information theory based feature selection scheme to select the most representative features that are sensitive to human motion. After that, we propose a crowd counting classifier based on transfer kernel learning and information fusion, which is robust to temporal and environmental disparities. We implemented WiFree with commodity WiFi routers and conducted extensive experiments in three indoor environments of disparate sizes. Experimental results have validated that WiFree is able to achieve 99.1% occupancy detection accuracy and 92.8% crowd counting accuracy over temporal variation in a device-free and privacy-preserving manner. (C) 2018 Elsevier B.V. All rights reserved.