vLLM Metal lets the vLLM serving engine run on a Mac. Upstream vLLM supplies the API server, scheduler and paged block manager; MLX and Metal run the model, so several requests can be served at once with a predictable memory budget.
It supports speculative decoding, GGUF checkpoints, KV cache offloading and speech-to-text, and has recipes by Mac memory size. Install it from the project's Homebrew tap, or with the install script for the latest development build.
It needs macOS 15 (Sequoia) or later on Apple silicon. It's free and open source (Apache 2.0) and runs on your Mac.
How to Get It
brew tap vllm-project/vllm-metal https://github.com/vllm-project/vllm-metal && brew install vllm-project/vllm-metal/vllm-metalcurl -fsSL https://raw.githubusercontent.com/vllm-project/vllm-metal/main/install.sh | bashOfficial downloads only. MLX never re-hosts an app, and never runs a command for you.
MLX Score
Ranked against the run models locally listings, from real public data only. How the score works.
Pairs Well With
Information
- Developer
- vLLM project
- Category
- Run models locally
- Price
- Free
- Your data
- Runs on your Mac
- Licence
- Apache-2.0
- Mac
- macOS 15.0 or later
- iPhone and iPad
- —
- Needs
- M1 or later
- Facts checked
- 2026-11-03
Sources
You Might Also Like
OllamaRun locally
llama.cppRun locally
LM StudioRun locally
Open WebUIRun locally
AnythingLLMRun locally
Locally AIRun locally
apfelRun locally
exoRun locally
oMLXRun locallyvLLM Metal is a product of vLLM project; MLX isn’t affiliated with it. Something wrong? Write to support@mlx.app.