basecompute / superfluid
git clone https://github.com/basecompute/superfluid.git
Superfluid is an open-source server that runs LLMs on your own hardware. It serves a model from one machine, or from several machines that act as one server. It supports llama.cpp, MLX, and BaseRT, and exposes OpenAI-, Anthropic-, and Ollama-compatible API endpoints.
+0
stars, 24h · #78,658
+~0
stars, 7 days · #78,980
+~0
stars, 30 days · #77,837
26
★ total
3
forks · 18 open issues
Star history (as recorded here)
LanguageRust
LicenseApache-2.0
Created2026-10-08 (1mo ago)
Last pushtoday
OwnerOrganization
GitHub topics
anthropic-apiapple-siliconclaude-codecoding-agentcontinuous-batchinginference-serverllama-cppllm-inferencellm-servinglocal-llmmlxollamaopenai-apirust
Links
- GitHub: github.com/basecompute/superfluid
- Homepage: https://superfluid.sh
- Releases · Issues · Hacker News mentions
README
Loading README…