Niko1221/Strata

Strata is a C++ inference engine enabling Qwen3.8-Flash-Next, a 125B parameter MoE model, to run on consumer hardware with 12-24 GB VRAM and 64 GB RAM. The project provides a one-click installer for Windows and Linux, exposing local OpenAI and Anthropic compatible APIs. This allows developers to test large-scale mixture-of-experts architectures on standard workstations without cloud dependency.