A single-header, zero-dependency (except the C++ STL) public domain implementation of the PLY mesh file format. An overview and definition of the file format is available here. This format is often used in the computer vision and graphics communities for its relative simplicity, ability to support arbitrary mesh attributes, and binary modes. Since ~2023 it has been used one of the primary interchange formats between applications & pipelines for gaussian splatting. Most famously, PLY is used to distribute models in the Stanford 3D Scanning Repository, including the original bunny.

The library is written in C++17 and requires a recent compiler. Tinyply supports importing and exporting PLY files in both binary and ascii formats. Tinyply supports filesizes >= 4gb and can read big-endian binary files (but not write them). Recently in the 3.0 release, tinyply adds a fast-path parser for the most common import scheme (binary little-endian) when specific conditions are met.

The project comes with a simple example program demonstrating a circular write / read and all of the major API functionality.

Since 2015, tinyply has been used in thousands of projects including pointcloud tools, raytracers, synthetic data renderers, computational geometry libraries, and more. It's even listed by the US Library of Congress as a utility for archival preservation of digital media. A few other notable projects are highlighted below:

• libigl, a robust computational geometry library from UoT professors Alec Jacobson and Daniele Panozzo.
• Maplab from ETH Zürich, a research-oriented visual-inertial mapping framework.
• glChAoS.P from Michele Morrone, a rendering sandbox for 3D strange attractors.
• Cilantro, a robust and featureful C++ library for working with pointcloud data.
• ScanNet, an RGB+D dataset of 2.5 million views across 1500 scans.
• PlaneRCNN, 3D plane detection via single-shot images from NVIDIA Research
• KNOSSOS, a framework to visualize and annotate 3D image data (neural morphology and connectivity).
• fVDB, a Python library for building applications using NanoVDB on the GPU in PyTorch.

tinyply not what you're looking for? tinyply trades some performance for simplicity and flexibility. For domain-specific uses (e.g. where your application does not need to handle arbitrary user-fed PLY files), there may be other alternatives. For more, please check out the following benchmarks:

• Vilya Harvey's ply-parsing-perf
• Maciej Halber's ply_io_benchmark

• version 2.0 is an API re-write to support later improvements towards variable length lists. One notable change is that tinyply now produces and consumes structured byte arrays, with type information held as metadata.
• version 2.1 contained minor bugfixes and speed improvements.
• version 2.2 is a rewrite of the inner read/write loop. Compared to version 2.0, this version reads and writes binary about five times faster. When a list size hint is given for reading, the performance is approximately comparable to rply.
• version 2.3 contains minor bugfixes and performance improvements.
• version 3.0 is a little less tiny, but it's a major performance upgrade and adds support for variable length lists alongside better error handling,

This software is in the public domain. Where that dedication is not recognized, you are granted a perpetual, irrevocable license to copy, distribute, and modify this file as you see fit. If these terms are not suitable to your organization, you may choose to license it under the terms of the 2-clause simplified BSD.