Open Source WebGPU Projects
WebGPU is a web standard that lets browser applications use a device’s graphics processing unit for rendering and general-purpose computation. It provides a modern alternative to older browser graphics APIs, with capabilities suited to detailed 3D visuals, simulations, image processing, and machine-learning workloads. By making GPU acceleration available through web technologies, WebGPU can support responsive, compute-intensive experiences without requiring users to install native software, though availability depends on browser, operating system, and hardware support.
Open source WebGPU tools include graphics libraries, rendering engines, compute frameworks, and utilities for running models in the browser. When choosing one, consider its maturity, license, maintenance activity, supported browsers and devices, performance needs, and how well it fits existing code and workflows. These tools can be useful to web developers, researchers, and teams exploring GPU-accelerated applications that run locally in a browser.
3 repositories · updated September 28, 2026

Shimmy: A Pure-Rust WebGPU Inference Engine for GGUF Models
Shimmy is a high-performance, pure-Rust WebGPU inference engine designed for GGUF models. It offers OpenAI-API compatibility, enabling local and private execution of large language models without Python or C++ dependencies. This single-binary solution provides rapid startup and a low memory footprint, making it an efficient alternative for local AI inference.

three.js: A Powerful JavaScript Library for 3D Graphics on the Web
three.js is a robust and easy-to-use JavaScript 3D library designed for rendering interactive 3D graphics directly in a web browser. It provides a comprehensive set of features for creating complex 3D scenes, animations, and effects, leveraging technologies like WebGL and WebGPU. Developers can quickly build immersive experiences for augmented reality, virtual reality, and general web applications.

whisper-web: Transcribe Speech in Your Browser
Whisper Web uses Transformers.js to run speech recognition in a browser. It suits people who want to try transcription without setting up a separate server, though browser support and performance may vary.