noumena-labs/tomegguf · webgpu · perf
the grimoire for local intelligence
COGENTLM
tome
A high-performance npm package for running local LLMs in the browser. GGUF native. WebGPU native. Magic in your tab. Shared GPU buffers and direct support for three.js.
◦ gguf native◦ webgpu native◦ zero server◦ three.js ready
stream open