Superfluid, an open-source local LLM server for multi-agent inference, debuts on the Hugging Face blog
Basecompute introduced Superfluid, an Apache-2.0 server for running LLMs on your own machine that speaks the OpenAI, Anthropic and Ollama APIs and can serve GGUF and MLX models straight from the Hugging Face Hub.
Photo: SiliconANGLE theCUBE / Wikimedia Commons, CC BY 3.0. Hugging Face logo via Wikimedia Commons (public domain).
Basecompute has introduced Superfluid, an open-source server for running large language models locally and built for multi-agent inference, in an October 8 post on the Hugging Face blog.
According to the post, Superfluid is released under the Apache-2.0 license and supports the OpenAI, Anthropic and Ollama APIs. It runs several back ends: llama.cpp for GGUF models, MLX for MLX directories and the company's BaseRT engine for .base bundles.
Users can serve a GGUF model directly from Hugging Face with a single command, choosing the quantization with a tag, the post shows.