Meshroom is a free, open-source, node-based visual programming framework designed for creating, managing, and executing complex data processing pipelines, with a particular focus on photogrammetry and 3D reconstruction. The latest version, Meshroom 2025.1.0, released on 18 August 2025, has evolved into a more flexible toolbox with a new plugin architecture. This architecture unifies standard computer vision pipelines, including photogrammetry, camera tracking, HDR panorama, Lidar Meshing, Raw image files conversion, and colour calibration, within the AliceVision plugin. The software also introduces new AI-powered capabilities, such as a semantic segmentation plugin, and offers a collection of open-source extensions available via the new MeshroomHub.
Meshroom is widely used in industries such as gaming, film, and virtual reality to produce high-quality 3D assets with minimal effort. By processing multiple images of an object or environment, it stitches them together to generate detailed 3D meshes that can be further refined or integrated into projects. The software is particularly valued for its ability to infer the geometry of a scene from a set of unordered photographs or videos, effectively reversing the projection of a 3D scene onto a 2D plane to recover depth information.
Availability and Cost
Meshroom is completely free and open-source, making it an attractive option for hobbyists, educators, and professionals who need a cost-effective 3D reconstruction tool. It is distributed under an open-source licence, and all prebuilt binaries are available for download at no cost. The software can be downloaded from several official sources, including the FileHorse website, the Meshroom official website, and the GitHub repository for AliceVision.
System Requirements and Compatibility
Meshroom is designed for Windows operating systems. The latest version, Meshroom 2025.1.0, is compatible with Windows 7, Windows 8, Windows 10, and Windows 11 (64-bit versions). Earlier versions, such as Meshroom 2023.2.0, also support these Windows versions. The software requires a quad-core CPU (Intel or AMD) and at least 16GB of RAM. For optimal performance, an NVIDIA GPU with CUDA capability is recommended; the latest binaries are built with CUDA-11 and are compatible with compute capability 3.5 or higher. A GPU such as the GTX 1070 or higher is suggested. Storage requirements include an SSD with at least 50GB of free space for large projects.
It is important to note that without a supported NVIDIA GPU, only the "Draft Meshing" option can be used for 3D reconstruction, which may limit the quality and speed of the output. On Linux distributions, such as Ubuntu, users may encounter conflicts between native drivers and Mesa drivers, which can result in an empty black window. For macOS, there is no official Meshroom release available, and users are directed to seek alternative information on the topic.
Key Features and Workflow
Meshroom employs a node-based visual programming interface. Each node represents a specific operation, and the output attributes of one node can seamlessly feed into subsequent steps. This system allows for a flexible and intuitive workflow where users can visually manage and adjust the reconstruction pipeline. When a node’s attribute is modified, only the affected downstream nodes are invalidated, while cached intermediate results are reused to minimise unnecessary computation.
The core functionality is centred on photogrammetry-based 3D reconstruction. Meshroom uses a powerful photogrammetry pipeline that converts 2D images into 3D models. This feature is particularly valuable for professionals needing accurate and realistic 3D representations of real-world objects. The process infers 3D geometry from 2D images, a technique that involves camera tracking and structure from motion. The software supports multiple file formats and includes interactive widgets for visualising images and 3D data.
Recent updates, as seen in the 2025.1.0 release, have introduced significant improvements. For instance, support for RAW image files has been added. Texturing quality has been largely improved, and texturing speed has seen enhancements. The texturing node now also supports UDIM. In the meshing stage, users can export the dense point cloud in Alembic format. A new option allows for the export of the full raw dense point cloud, which includes all 3D points candidates before cut and filtering. MeshFiltering now correctly preserves colour data, and there is an option to export colour data per vertex.
Other notable features include a new node for semantic image segmentation, support for pixel aspect ratio, noise reduction in HDR merging, and a 2D viewer with an image sequence player. The StructureFromMotion node now supports automatic alignment of the 3D reconstruction. A new node for intrinsics and rig calibration using a multiview acquisition of a checkerboard has been introduced, along with a new Nodal Camera Tracking pipeline. The software also manages LCP (Lens Correction Profile) in the imageProcessing node and adds a slider in the Viewer3D to display cameras based on their resection IDs.
Performance and Processing
Meshroom supports both local and distributed execution, enabling efficient parallel processing on render farms. This capability is crucial for handling large datasets and complex reconstructions in a timely manner. However, processing can be slow without proper hardware, particularly without a supported NVIDIA GPU for CUDA acceleration. The software is designed to work with high-quality images and proper configuration to produce highly accurate models.
Suitability for Different Users
Meshroom is suitable for beginners due to its intuitive interface. However, some understanding of photogrammetry and 3D modeling principles can be beneficial to achieve the best results. For professionals who require more control over the reconstruction process, an alternative such as Agisoft Metashape is available, which offers robust features and extensive support for various workflows.
Development and Community
Meshroom is developed by AliceVision, an open-source project. The software has an active development community, as evidenced by recent events such as a presentation by Candice Bentéjac and Fabien Castan at an open-source after-work event in December 2025. Furthermore, a PhD defence was dedicated to improving the robustness and accuracy of 3D reconstruction methods by integrating explicit models and leveraging physical and geometric properties, indicating ongoing academic and practical research in the field.
Conclusion
Meshroom 2025.1.0 represents a significant evolution in free, open-source tools for 3D reconstruction. Its node-based visual programming framework, combined with a new plugin architecture and AI-powered capabilities, offers a flexible and powerful solution for creating high-quality 3D models from 2D images. While it is optimally performed on Windows with an NVIDIA GPU, its cost-free nature makes it accessible to a wide range of users, from hobbyists to professionals in gaming, film, and virtual reality. The software's continuous development and active community support further enhance its value as a tool for photogrammetry and computer vision tasks.