Context-aware computers rely on user and physical models to describe the context of a user. In this paper, we focus on the problem of developing and maintaining a physical model of the environment using a mobile computer. We describe a set of tools for automatically creating and modifying three-dimensional contextual information. The tools can be utilized across multiple hardware platforms with different capabilities, and operating in collaboration with one another. We demonstrate the capabilities of the tools using two mobile platforms. One of them, a mobile augmented reality system is used to contruct a geometric model of an indoor environment which is then visualized on the same platform.