OpenAI-compatible LLM serving stack for FuriosaAI RNGD accelerators
furiosa-llm is FuriosaAI's inference serving software for RNGD hardware. Installed via pip, it serves pre-compiled models behind an OpenAI-compatible endpoint and a Python API, so existing clients can point at a new base URL without rewriting code. It is part of the Furiosa SDK and documentation, which also cover deployment and lower-level programming of the accelerators.
For people
For agents
Nothing listed yet.