glTF2Image

glTF2Image is a package for rendering screenshots or thumbnail images of 3D models. It provides an easy-to-use .NET API around Filament and SwiftShader. It is suitable for web servers that don't have a GPU.

Example

// Create a Renderer instance.
await using var renderer = await Renderer.CreateAsync();

// Load the model.
await using var model = renderer.CreateGLTFAsset(File.ReadAllBytes(Path.Join(TestDataPath, "Avocado.glb")));

// Load another gltf model defining the lights and camera.
await using var lightsAndCamera = renderer.CreateGLTFAsset(File.ReadAllBytes(Path.Join(TestDataPath, "avocado_lights_and_camera.gltf")));

// Render the scene.
int width = 576;
int height = 324;
var data = await renderer.RenderAsync(576, 324, new[] { model, lightsAndCamera });

// We currently have a pixel buffer in RGBA format. Use your favorite library to encode this as a PNG.
// For this example, we'll use ImageSharp.
var image = Image.LoadPixelData<Rgba32>(data, width, height);
await image.SaveAsPngAsync("avocado.png");

Result:

Preparing your model

glTF2Image accepts models in glTF or glb (glTF binary) format.

Your scene must have exactly one camera to tell glTF2Image what to render. If your model does not define a camera, you can add a separate glTF file with a camera, like the example above.

To validate your model, download a recent Filament release and use the gltf_viewer tool to ensure that Filament renders your model as expected. Be sure to uncheck the "Scene > Scale to unit cube" option, since glTF2Image will not scale your model.

Filament has good support for many glTF features and extensions. There are a small number of limitations:

• Only one directional light is supported

Requirements

• .NET 8 or later
• Windows or Linux
• x64