This guide illustrates the process of developing a 3D object viewer tailored for WebAssembly. Notably, it offers the flexibility to select the rendering backend, either OpenGL or WebGPU, during runtime.

The additional steps required are:

1. Update CMakeLists.txt so that the VTK::RenderingWebGPU component is requested when finding VTK.
2. Link your targets with RenderingWebGPU. There is no problem linking to both RenderingWebGPU and RenderingOpenGL2.
3. Pass "--vtk-factory-prefer RenderingBackend=WebGPU" as command line arguments when initializing the VTK WASM module to select the WebGPU backend at runtime.
4. Call vtkObjectFactory::InitializePreferencesFromCommandLineArgs(argc, argv) in your main function before creating any VTK object. This will ensure that the rendering backend preference is set based on the command line arguments.

emcmake cmake \
  -G Ninja \
  -S /path/to/vtk/Examples/Emscripten/Cxx/ConeMultiBackend \
  -B out/build \
  -DVTK_DIR=/path/to/where/vtk/wasm/was/built

cmake --build out/build

# Configure
docker run --rm -it \
  -v"$PWD":/work kitware/vtk-wasm-sdk:latest \
  emcmake cmake -GNinja -S /work -B /work/build -DVTK_DIR=/VTK-install/Release/wasm32/lib/cmake/vtk
# Build
docker run --rm -it \
  -v"$PWD":/work kitware/vtk-wasm-sdk:latest \
  cmake --build /work/build

docker run --rm -it `
  -v"$PWD":/work kitware/vtk-wasm-sdk:latest `
  emcmake cmake -GNinja -S /work -B /work/build -DVTK_DIR=/VTK-install/Release/wasm32/lib/cmake/vtk
# Build
docker run --rm -it `
  -v"$PWD":/work kitware/vtk-wasm-sdk:latest `
  cmake --build /work/build

cd out/build
python3 -m http.server 8000

Open your browser to http://localhost:8000