Show HN: Janus – Go binary that runs GGUF models via Vulkan on AMD/Intel/Nvidia

Janus packages local LLM inference into a single Go binary with an OpenAI-compatible API, letting teams run GGUF models on AMD, Intel, or Nvidia GPUs—or CPU—without Python, Docker, or a cloud dependency. For CIOs, this can reduce recurring inference costs, improve data control, and accelerate private AI deployments that still plug into existing OpenAI-based tools and workflows. IT organizations should view it as a lightweight path to on-prem or edge AI, but one that requires disciplined model management, GPU/driver standardization, and operational guardrails to avoid fragmentation.