Open3D

Overview

Open3D is a free, open-source toolkit for building software around 3D data, best suited to developers and research teams working with point clouds, meshes, or 3D machine learning. Its breadth is a strength, but it is a programming library rather than a ready-made scanning workflow, and some platform and remote-viewing caveats matter.

C++ and Python interfaces expose data structures and processing algorithms for reconstruction, alignment, and visualization. The project also provides a standalone viewer and web visualization, while the MIT license permits research and commercial use with proper attribution.

Supported inputs include Intel RealSense cameras, Azure Kinect, RGB-D sequences, RealSense bag files, point-cloud files, and mesh files. Export formats include XYZ, XYZN, XYZRGB, PTS, PLY, PCD, STL, OBJ, OFF, GLTF, GLB, and FBX, giving it room in workflows that move between point clouds and common mesh formats.

Key features

Processing and visualization

Core tools cover 3D data structures, processing algorithms, scene reconstruction, and surface alignment. Physically based rendering supports inspection of results, and GPU acceleration is listed for core operations; the v0.20 release announcement highlights full Intel GPU support. This combination suits teams building their own processing pipeline, though it asks more of users than a single-purpose desktop scanner app.

Machine learning

Open3D-ML adds semantic point-cloud segmentation, pretrained models, and training pipelines. It supports PyTorch and TensorFlow and includes framework-independent visualization, making it useful when 3D processing needs to meet model training or inference. There is an important Linux exception: since v0.18, the PyPI Linux wheel does not provide native TensorFlow support because of PyTorch/TensorFlow build incompatibilities.

Viewer and browser visualization

Open3D-Viewer is a standalone app for Debian and Ubuntu, macOS, and Windows. The web visualizer can run standalone C++ or Python servers and display output in a browser through WebRTC; it can also be embedded in Python Jupyter Notebook or JupyterLab cells. That makes it a flexible route for sharing or notebook-based inspection, but the web visualizer uses HTTP by default. Sensitive remote visualization calls for appropriate web-server configuration and SSL certificates.

Build and supply-chain choices

Current Python packages support Ubuntu 22.04+, macOS 10.15+, and 64-bit Windows 10+, with Python 3.10 through 3.14. The installation guide offers standard, CPU-only, and Intel GPU wheels, plus a preview Windows CUDA wheel; users on other systems can build from source. Release Python wheels and C++ binaries carry signed SLSA build-provenance attestations, a useful supply-chain signal for teams incorporating dependencies into larger deployments.

Pricing

Open3D is free, with no free trial because there is no paid plan to evaluate. The free offering is the library itself, under the MIT license, with C++ and Python interfaces, the viewer, and Open3D-ML. It is a strong fit for individuals and organizations able to work with a developer toolkit; no paid tier or seat-based plan is part of the offer.

Platforms

Open3D supports Linux, macOS, and Windows, with API access and self-hosted and web visualization options. The standalone viewer runs on Debian/Ubuntu, macOS, and Windows. Python wheel support is narrower than the broad platform labels: Ubuntu 22.04+, macOS 10.15+, and 64-bit Windows 10+ are supported by current packages, while other systems require a source build. Linux users who need TensorFlow through the PyPI wheel should account for the native-support limitation.

Who it's for

Open3D is a good match for software developers, research projects, and cloud deployments that need programmable 3D processing, capture, reconstruction, or machine-learning tools. Its support for camera inputs, point clouds, meshes, and a wide range of export formats gives it utility across stages of a 3D data workflow. It is less suitable for someone seeking a polished, scanner-specific workflow without programming, or for teams that need to visualize sensitive remote data without configuring secure web serving.

For bugs and feature requests, the project directs users to GitHub Issues; usage discussions are served by a forum, with Discord for chat and collaboration. The project also welcomes open-source contributions.

Pros and cons

• Pro: C++ and Python APIs, processing tools, reconstruction, and visualization live in one open-source toolkit, useful for teams building a custom 3D pipeline.
• Pro: Open3D-ML adds segmentation, pretrained models, and training pipelines, extending the library beyond geometry processing.
• Pro: Broad input and export format support helps connect camera capture, point clouds, and mesh workflows.
• Pro: MIT licensing allows research and commercial use with attribution, and release binaries include signed provenance attestations.
• Con: The Linux PyPI wheel lacks native TensorFlow support from v0.18 onward, complicating that framework choice for Linux users.
• Con: The web visualizer defaults to HTTP, so sensitive remote viewing requires additional SSL and server configuration.
• Con: Current Python wheels target specific operating-system versions; unsupported systems need a source build.

Alternatives

Choose MeshInspector instead if you want a freemium, cross-platform app centered on viewing and basic editing of meshes, point clouds, and voxels, with basic mesh repair. Choose JMStudio for 3DMakerpro scanner workflows and its STL, OBJ, PLY, ASC, and RSCAN exports. Revo Scan 6 is a free option for scanning and post-processing with Revopoint 3D scanners.

MeshLab is another free, open-source mesh processing and editing system. Polycam has a free mobile and web plan with limited space, object, and AI captures, GLTF exports, and public link sharing; its paid Basic plan is $150 a year. RealityScan is a freemium alternative for individuals and small businesses under $1 million USD in revenue in the past 12 months, educational institutions, and students.

Polyga PointKit offers a Premium plan at $99 a month, billed 1 month, for XTract3D 2 and PointKit with a single active license. Structure offers an SDK Developer License with a 7 day free trial and all SDK features for experimentation and prototyping.

Browse more options in 3D Scanning Software and Point Cloud Software.

Verdict

Choose Open3D if you need a free, extensible foundation for 3D processing, visualization, or machine-learning work and can accommodate its development-oriented workflow. Its strongest reason to choose it is the combination of broad APIs and capable 3D tooling under an MIT license; look elsewhere if you need an off-the-shelf scanner app, Linux TensorFlow through the PyPI wheel, or secure remote visualization without extra configuration.