Intravital microscopy (IVM) enables live imaging of animals at single-cell resolution, offering essential insights into cancer progression. This technique allows for the observation of single-cell behaviors within their natural 3D tissue environments, shedding light on how genetic and microenvironmental changes influence the complex dynamics of tumors. IVM generates highly complex datasets that often exceed the analytical capacity of traditional uni-parametric approaches, which can neglect single-cell heterogeneous in vivo behavior and limit insights into microenvironmental influences on cellular behavior. To overcome these limitations, we present BEHAV3D Tumor Profiler (BEHAV3D-TP), a computational framework that enables unbiased single-cell classification based on a range of morphological and dynamic single cell features. BEHAV3D-TP is user-friendly and integrates with widely used 2D and 3D image processing pipelines, enabling researchers to profile cancer and healthy cell dynamics in IVM data without advanced computational expertise. Here, we apply BEHAV3D-TP to study diffuse midline glioma (DMG), a highly aggressive pediatric brain tumor characterized by invasive progression. By extending BEHAV3D-TP to incorporate tumor microenvironment (TME) data from IVM or fixed correlative imaging, we demonstrate that distinct migratory behaviors of DMG cells are associated with specific TME components, including tumor-associated macrophages and vasculature. BEHAV3D-TP enhances the accessibility of computational tools for analyzing the complex behaviors of cancer cells and their interactions with the TME in IVM data.