Devices, methods, and systems for provisioning network credentials onto a device using audible commands. In some instances, the device is a headless device without a graphical user interface. The device may include a limited number of speech models and phrases pre-programmed onto the device to enable the device to query a user to enter network credentials and have a dialog type interface with the user using local speech recognition techniques. Once the network credentials are entered and the device connects to the user's network, the device may change modes and limit the local speech recognition techniques. In this respect, the device may send audio input to a more powerful remote device for processing.