Project Maya enables private run of GLM-5.3-Flash (321B MoE) across GPUs from one to 16, with an engine, server, dashboard, and OpenAI- and Anthropic-compatible API on Linux or Windows (NVIDIA or AMD).
Read the original at github.com→★ 796 stars
Original headline: "mw00/project-maya: Project Maya - run GLM-5.3-Flash (321B MoE) privately on your own GPU(s), from one card to 16, NVIDIA or AMD, Linux or Windows. Engine, server, dashboard, OpenAI- and Anthropic-compatible API, quant tools."