mlx-serve

(★ 474)

Native LLM inference server for Apple Silicon. OpenAI + Anthropic API compatible. No Python. Includes MLX Core macOS app with chat, agent mode, and tool calling.

mlx-serve Latest Version Download

# Use via CDN

jsDelivr

jsDelivr serves any public GitHub repository as a CDN with zero setup. Pick a version and a file to get a ready-to-paste link and snippet.

Command Glossary

Commands referenced in this DOCs, explained below.

🔍

brew install

View Details ▼

Install a Homebrew formula or cask.

brew install {{formula|cask}}

Install a formula/cask:

brew install {{[-s|--build-from-source]}} {{formula}}

Build and install a formula from source (dependencies will still be installed from bottles):

brew install {{[-n|--dry-run]}} {{formula|cask}}

Download the manifest, print what would be installed but don't actually install anything:

🔍

brew tap

View Details ▼

Tap a Homebrew formula repository. If no arguments are provided, list all installed Homebrew taps.

brew tap

List installed Homebrew taps:

brew tap {{github_username}}/{{github_repository_name}}

Tap a Git repository hosted on GitHub:

brew tap {{username}}/{{repository_name}} {{git_clone_url}}

Tap a Git repository hosted outside of GitHub:

🔍

brew upgrade

View Details ▼

Upgrade outdated formulae and casks.

brew upgrade

Upgrade all outdated casks and formulae:

brew upgrade {{formula|cask}}

Upgrade a specific formula/cask:

brew upgrade {{[-n|--dry-run]}}

Print what would be upgraded, but don't actually upgrade anything:

🔍

git clone

View Details ▼

Clone an existing repository.

git clone {{remote_repository_location}} {{path/to/directory}}

Clone an existing repository into a new directory (the default directory is the repository name):

git clone --recursive {{remote_repository_location}}

Clone an existing repository and its submodules:

git clone {{[-n|--no-checkout]}} {{remote_repository_location}}

Clone only the `.git` directory of an existing repository:

🔍

Hello

View Details ▼

Print "Hello, world!", "hello, world", or a customizable text.

hello

Print "Hello, world!":

hello {{[-t|--traditional]}}

Print "hello, world", the traditional type:

hello {{[-g|--greeting]}} "{{greeting_text}}"

Print a text message:

🔍

pip

View Details ▼

Python package manager.
Some subcommands such as `install` have their own usage documentation.

pip install {{package}}

Install a package (see `pip install` for more install examples):

pip install --user {{package}}

Install a package to the user's directory instead of the system-wide default location:

pip install {{[-U|--upgrade]}} {{package}}

Upgrade a package:

🔍

serve

View Details ▼

Static file serving and directory listing.

serve

Start an HTTP server listening on the default port to serve the current directory:

serve -p {{port}} {{path/to/directory}}

Start an HTTP server on a specific [p]ort to serve a specific directory:

serve {{[-C|--cors]}}

Start an HTTP server with CORS enabled by including the `Access-Control-Allow-Origin: *` header in all responses:

🔍

zig

View Details ▼

The Zig compiler and toolchain.

zig build

Compile the project in the current directory:

zig build run

Compile and run the project in the current directory:

zig init

Initialize a `zig build` project with library and executable:

🔍

gemma3

View Details ▼

The examples for this command have been moved together with `inference-snap`.

tldr inference-snap

View documentation for `inference-snap`:

🔍

gemma4

View Details ▼

The examples for this command have been moved together with `inference-snap`.

tldr inference-snap

View documentation for `inference-snap`:

🔍

prompt

View Details ▼

Change the default DOS style prompt in a command window.

prompt

Reset the prompt to the default setting:

prompt {{prompt}}

Set a specific prompt:

prompt $D $P$G

Change the prompt to show the current date first:

// repository documentation