Recognizing the crucial role of mechanical regulation and forces in tissue development and homeostasis has stirred a demand for in situ measurement of forces and stresses. Among emerging techniques, the use of cell geometry to infer cell junction tensions, cell pressures and tissue stress has gained popularity owing to the development of computational analyses. This approach is non-destructive and fast, and statistically validated based on comparisons with other techniques. However, its qualitative and quantitative limitations, in theory as well as in practice, should be examined with care. In this Primer, we summarize the underlying principles and assumptions behind stress inference, discuss its validity criteria and provide guidance to help beginners make the appropriate choice of its variants. We extend our discussion from two-dimensional stress inference to three dimensional, using the early mouse embryo as an example, and list a few possible extensions. We hope to make stress inference more accessible to the scientific community and trigger a broader interest in using this technique to study mechanics in development.