Enhance the understanding of 3D reconstruction capabilities from images and photogrammetry by creating a C++ framework. Facilitate reproducible research through an easily comprehensible and precise implementation of both contemporary and traditional algorithms. OpenMVG is crafted to be user-friendly, allowing for straightforward reading, learning, modification, and application. With its rigorous test-driven development and comprehensive samples, the library empowers users to construct reliable larger systems. OpenMVG encompasses a complete framework for 3D reconstruction from images, comprising various libraries, executables, and processing pipelines. The libraries grant seamless access to functionalities such as image manipulation, feature description and matching, camera models, feature tracking, robust estimation, and multiple-view geometry. The executables address specific tasks required by a pipeline, including scene initialization, feature detection, matching, and the reconstruction process known as structure-from-motion. Furthermore, this versatility makes OpenMVG a valuable tool for both beginners and seasoned researchers in the field.