About DeepShader

DeepShader is a deep learning framework focused on computer vision, built for real-time visualization of every processing step — from training to inference — using WebGPU.

It combines low-level GPU programming with web portability: you define execution graphs between compute shaders and GPU resources, all linked to canvases for constant visual feedback on what your model is actually doing.

Why DeepShader

• Transparency — see every texture and buffer update as it happens, instead of debugging blind.
• Performance — compute shaders run directly on the GPU through WebGPU, no server round-trip required.
• Portability — everything runs in the browser, so a project can be shared with a single link.

Get involved

DeepShader is under active development. Head over to the documentation to get started, or explore projects shared by the community from the home page.