draco repository

Repository Summary

Checkout URI https://github.com/google/draco.git
VCS Type git
VCS Version 1.3.6
Last Updated 2020-03-03
CI status No Continuous Integration
Tags No category tags.
Contributing Help Wanted (0)
Good First Issues (0)
Pull Requests to Review (0)


No packages found.



Version 1.3.6 release

  • WASM and JavaScript decoders are now hosted from a static URL
    • It is recommended to always pull your Draco WASM and JavaScript decoders from this URL:
    • https://www.gstatic.com/draco/v1/decoders/
    • Users will benefit from having the Draco decoder in cache as more sites start using the static URL
  • Changed web examples to pull Draco decoders from static URL
  • Added new API to Draco WASM decoder, which increased performance by ~15%
  • Decreased Draco WASM decoder size by ~20%
  • Added support for generic and multiple attributes to Draco Unity plug-ins
  • Added new API to Draco Unity, which increased decoder performance by ~15%
  • Changed quantization defaults:
    • POSITION: 11
    • NORMAL: 7
    • TEX_COORD: 10
    • COLOR: 8
    • GENERIC: 8
  • Code cleanup
  • Bug fixes

Version 1.3.5 release

  • Added option to build Draco for Universal Scene Description
  • Code cleanup
  • Bug fixes

Version 1.3.4 release

  • Released Draco Animation code
  • Fixes for Unity
  • Various file location and name changes

Version 1.3.3 release

  • Added ExpertEncoder to the Javascript API
    • Allows developers to set quantization options per attribute id
  • Bug fixes

Version 1.3.2 release

  • Bug fixes

Version 1.3.1 release

  • Fix issue with multiple attributes when skipping an attribute transform

Version 1.3.0 release

  • Improved kD-tree based point cloud encoding
    • Now applicable to point clouds with any number of attributes
    • Support for all integer attribute types and quantized floating point types
  • Improved mesh compression up to 10% (on average ~2%)
    • For meshes, the 1.3.0 bitstream is fully compatible with 1.2.x decoders
  • Improved Javascript API
    • Added support for all signed and unsigned integer types
    • Added support for point clouds to our Javascript encoder API
  • Added support for integer properties to the PLY decoder
  • Bug fixes

Previous releases



Draco is a library for compressing and decompressing 3D geometric meshes and point clouds. It is intended to improve the storage and transmission of 3D graphics.

Draco was designed and built for compression efficiency and speed. The code supports compressing points, connectivity information, texture coordinates, color information, normals, and any other generic attributes associated with geometry. With Draco, applications using 3D graphics can be significantly smaller without compromising visual fidelity. For users, this means apps can now be downloaded faster, 3D graphics in the browser can load quicker, and VR and AR scenes can now be transmitted with a fraction of the bandwidth and rendered quickly.

Draco is released as C++ source code that can be used to compress 3D graphics as well as C++ and Javascript decoders for the encoded data.



See BUILDING for building instructions.



For the best information about using Unity with Draco please visit https://gitlab.com/atteneder/DracoUnity

For a simple example of using Unity with Draco see README in the unity folder.

WASM and JavaScript Decoders

It is recommended to always pull your Draco WASM and JavaScript decoders from:


Users will benefit from having the Draco decoder in cache as more sites start using the static URL.

Command Line Applications

The default target created from the build files will be the draco_encoder and draco_decoder command line applications. For both applications, if you run them without any arguments or -h, the applications will output usage and options.

Encoding Tool

draco_encoder will read OBJ or PLY files as input, and output Draco-encoded files. We have included Stanford's Bunny mesh for testing. The basic command line looks like this:

./draco_encoder -i testdata/bun_zipper.ply -o out.drc

A value of 0 for the quantization parameter will not perform any quantization on the specified attribute. Any value other than 0 will quantize the input values for the specified attribute to that number of bits. For example:

./draco_encoder -i testdata/bun_zipper.ply -o out.drc -qp 14

will quantize the positions to 14 bits (default is 11 for the position coordinates).

In general, the more you quantize your attributes the better compression rate you will get. It is up to your project to decide how much deviation it will tolerate. In general, most projects can set quantization values of about 11 without any noticeable difference in quality.

The compression level (-cl) parameter turns on/off different compression features.

./draco_encoder -i testdata/bun_zipper.ply -o out.drc -cl 8

In general, the highest setting, 10, will have the most compression but worst decompression speed. 0 will have the least compression, but best decompression speed. The default setting is 7.

Encoding Point Clouds

You can encode point cloud data with draco_encoder by specifying the -point_cloud parameter. If you specify the -point_cloud parameter with a mesh input file, draco_encoder will ignore the connectivity data and encode the positions from the mesh file.

./draco_encoder -point_cloud -i testdata/bun_zipper.ply -o out.drc

This command line will encode the mesh input as a point cloud, even though the input might not produce compression that is representative of other point clouds. Specifically, one can expect much better compression rates for larger and denser point clouds.

Decoding Tool

draco_decoder will read Draco files as input, and output OBJ or PLY files. The basic command line looks like this:

./draco_decoder -i in.drc -o out.obj

C++ Decoder API

If you'd like to add decoding to your applications you will need to include the draco_dec library. In order to use the Draco decoder you need to initialize a DecoderBuffer with the compressed data. Then call DecodeMeshFromBuffer() to return a decoded mesh object or call DecodePointCloudFromBuffer() to return a decoded PointCloud object. For example:

draco::DecoderBuffer buffer;
buffer.Init(data.data(), data.size());

const draco::EncodedGeometryType geom_type =
if (geom_type == draco::TRIANGULAR_MESH) {
  unique_ptr<draco::Mesh> mesh = draco::DecodeMeshFromBuffer(&buffer);
} else if (geom_type == draco::POINT_CLOUD) {
  unique_ptr<draco::PointCloud> pc = draco::DecodePointCloudFromBuffer(&buffer);

Please see src/draco/mesh/mesh.h for the full Mesh class interface and src/draco/point_cloud/point_cloud.h for the full PointCloud class interface.

Javascript Encoder API

The Javascript encoder is located in javascript/draco_encoder.js. The encoder API can be used to compress mesh and point cloud. In order to use the encoder, you need to first create an instance of DracoEncoderModule. Then use this instance to create MeshBuilder and Encoder objects. MeshBuilder is used to construct a mesh from geometry data that could be later compressed by Encoder. First create a mesh object using new encoderModule.Mesh() . Then, use AddFacesToMesh() to add indices to the mesh and use AddFloatAttributeToMesh() to add attribute data to the mesh, e.g. position, normal, color and texture coordinates. After a mesh is constructed, you could then use EncodeMeshToDracoBuffer() to compress the mesh. For example:

const mesh = {
  indices : new Uint32Array(indices),
  vertices : new Float32Array(vertices),
  normals : new Float32Array(normals)

const encoderModule = DracoEncoderModule();
const encoder = new encoderModule.Encoder();
const meshBuilder = new encoderModule.MeshBuilder();
const dracoMesh = new encoderModule.Mesh();

const numFaces = mesh.indices.length / 3;
const numPoints = mesh.vertices.length;
meshBuilder.AddFacesToMesh(dracoMesh, numFaces, mesh.indices);

meshBuilder.AddFloatAttributeToMesh(dracoMesh, encoderModule.POSITION,
  numPoints, 3, mesh.vertices);
if (mesh.hasOwnProperty('normals')) {
    dracoMesh, encoderModule.NORMAL, numPoints, 3, mesh.normals);
if (mesh.hasOwnProperty('colors')) {
    dracoMesh, encoderModule.COLOR, numPoints, 3, mesh.colors);
if (mesh.hasOwnProperty('texcoords')) {
    dracoMesh, encoderModule.TEX_COORD, numPoints, 3, mesh.texcoords);

if (method === "edgebreaker") {
} else if (method === "sequential") {

const encodedData = new encoderModule.DracoInt8Array();
// Use default encoding setting.
const encodedLen = encoder.EncodeMeshToDracoBuffer(dracoMesh,

Please see src/draco/javascript/emscripten/draco_web_encoder.idl for the full API.

Javascript Decoder API

The Javascript decoder is located in javascript/draco_decoder.js. The Javascript decoder can decode mesh and point cloud. In order to use the decoder, you must first create an instance of DracoDecoderModule. The instance is then used to create DecoderBuffer and Decoder objects. Set the encoded data in the DecoderBuffer. Then call GetEncodedGeometryType() to identify the type of geometry, e.g. mesh or point cloud. Then call either DecodeBufferToMesh() or DecodeBufferToPointCloud(), which will return a Mesh object or a point cloud. For example:

// Create the Draco decoder.
const decoderModule = DracoDecoderModule();
const buffer = new decoderModule.DecoderBuffer();
buffer.Init(byteArray, byteArray.length);

// Create a buffer to hold the encoded data.
const decoder = new decoderModule.Decoder();
const geometryType = decoder.GetEncodedGeometryType(buffer);

// Decode the encoded geometry.
let outputGeometry;
let status;
if (geometryType == decoderModule.TRIANGULAR_MESH) {
  outputGeometry = new decoderModule.Mesh();
  status = decoder.DecodeBufferToMesh(buffer, outputGeometry);
} else {
  outputGeometry = new decoderModule.PointCloud();
  status = decoder.DecodeBufferToPointCloud(buffer, outputGeometry);

// You must explicitly delete objects created from the DracoDecoderModule
// or Decoder.

Please see src/draco/javascript/emscripten/draco_web_decoder.idl for the full API.

Javascript Decoder Performance

The Javascript decoder is built with dynamic memory. This will let the decoder work with all of the compressed data. But this option is not the fastest. Pre-allocating the memory sees about a 2x decoder speed improvement. If you know all of your project's memory requirements, you can turn on static memory by changing CMakeLists.txt accordingly.

Metadata API

Starting from v1.0, Draco provides metadata functionality for encoding data other than geometry. It could be used to encode any custom data along with the geometry. For example, we can enable metadata functionality to encode the name of attributes, name of sub-objects and customized information. For one mesh and point cloud, it can have one top-level geometry metadata class. The top-level metadata then can have hierarchical metadata. Other than that, the top-level metadata can have metadata for each attribute which is called attribute metadata. The attribute metadata should be initialized with the correspondent attribute id within the mesh. The metadata API is provided both in C++ and Javascript. For example, to add metadata in C++:

draco::PointCloud pc;
// Add metadata for the geometry.
std::unique_ptr<draco::GeometryMetadata> metadata =
  std::unique_ptr<draco::GeometryMetadata>(new draco::GeometryMetadata());
metadata->AddEntryString("description", "This is an example.");

// Add metadata for attributes.
draco::GeometryAttribute pos_att;
pos_att.Init(draco::GeometryAttribute::POSITION, nullptr, 3,
             draco::DT_FLOAT32, false, 12, 0);
const uint32_t pos_att_id = pc.AddAttribute(pos_att, false, 0);

std::unique_ptr<draco::AttributeMetadata> pos_metadata =
        new draco::AttributeMetadata(pos_att_id));
pos_metadata->AddEntryString("name", "position");

// Directly add attribute metadata to geometry.
// You can do this without explicitly add |GeometryMetadata| to mesh.
pc.AddAttributeMetadata(pos_att_id, std::move(pos_metadata));

To read metadata from a geometry in C++:

// Get metadata for the geometry.
const draco::GeometryMetadata *pc_metadata = pc.GetMetadata();

// Request metadata for a specific attribute.
const draco::AttributeMetadata *requested_pos_metadata =
  pc.GetAttributeMetadataByStringEntry("name", "position");

Please see src/draco/metadata and src/draco/point_cloud for the full API.

NPM Package

Draco NPM NodeJS package is located in javascript/npm/draco3d. Please see the doc in the folder for detailed usage.

three.js Renderer Example

Here's an example of a geometric compressed with Draco loaded via a Javascript decoder using the three.js renderer.

Please see the javascript/example/README.md file for more information.


For questions/comments please email draco-3d-discuss@googlegroups.com

If you have found an error in this library, please file an issue at https://github.com/google/draco/issues

Patches are encouraged, and may be submitted by forking this project and submitting a pull request through GitHub. See CONTRIBUTING for more detail.


Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at


Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.


Bunny model from Stanford's graphic department https://graphics.stanford.edu/data/3Dscanrep/


Want to contribute? Great! First, read this page (including the small print at the end).

Before you contribute

Before we can use your code, you must sign the Google Individual Contributor License Agreement (CLA), which you can do online. The CLA is necessary mainly because you own the copyright to your changes, even after your contribution becomes part of our codebase, so we need your permission to use and distribute your code. We also need to be sure of various other thingsā€”for instance that you'll tell us if you know that your code infringes on other people's patents. You don't have to sign the CLA until after you've submitted your code for review and a member has approved it, but you must do it before we can put your code into our codebase. Before you start working on a larger contribution, you should get in touch with us first through the issue tracker with your idea so that we can help out and possibly guide you. Coordinating up front makes it much easier to avoid frustration later on.

Code reviews

All submissions, including submissions by project members, require review. We use GitHub pull requests for this purpose. Please make sure that your code conforms with our coding style guidelines.

The small print

Contributions made by corporations are covered by a different agreement than the one above, the Software Grant and Corporate Contributor License Agreement.