Installation

You need an Apple Silicon Mac (MLX is Metal-only, so macOS only by design). There are four ways in — pick whichever fits.

Homebrew (recommended)

A self-contained, signed + notarized binary — no toolchain to install by hand:

brew install joshuarossi/tap/mlx-bun
mlx-bun

That’s the whole thing. The bottle already bundles the MLX native runtime, so the only thing the first run fetches is the model. Upgrade later with brew upgrade mlx-bun.

Direct download

No Homebrew, Bun, or git — just download the signed, notarized binary:

install script

curl -fsSL https://mlx-bun.dev/install.sh | sh
mlx-bun

tarball

mkdir mlx-bun && cd mlx-bun
curl -fsSL https://github.com/joshuarossi/mlx-bun/releases/latest/download/mlx-bun-arm64.tar.gz | tar xz
./mlx-bun

The bundle is self-contained (binary + MLX runtime) and notarized, so it runs even on a browser download — no Gatekeeper prompt. The tarball variant leaves a self-contained folder; symlink ./mlx-bun into your PATH to keep it.

bunx (no install)

Already have Bun ≥ 1.3.14? Run it straight from npm, nothing to install:

bunx mlx-bun

First run fetches the MLX native runtime (~52 MB) and the model into your caches, then serves.

Note

mlx-bun is Bun-only by design — npx mlx-bun under Node exits with a “requires Bun” notice. Use bunx.

From source

For development, or to use the library from a clone. Needs Bun ≥ 1.3.14 — no Python, no native library to install by hand:

# 1. Install Bun
curl -fsSL https://bun.sh/install | bash
exec $SHELL -l

# 2. Clone, install, link the CLI
git clone https://github.com/joshuarossi/mlx-bun.git && cd mlx-bun
bun install
bun run link-cli      # adds the `mlx-bun` command — or run `bun src/cli.ts <verb>`

# 3. Run it
mlx-bun

What happens on first run

Bare mlx-bun is an alias for mlx-bun serve. With no model named, it does everything for you:

1. pulls the MLX native runtime (~52 MB) into ~/Library/Caches/mlx-bun/ (skipped on the Homebrew/direct-download installs — the runtime ships in the bundle);
2. downloads the sub-GB MiniCPM5-1B starter and serves it, so you’re chatting in well under a minute;
3. starts downloading gemma-4-e4b (the stronger 4B model) in the background — it becomes the default on your next mlx-bun serve, on a 16 GB+ Mac;
4. opens the chat UI in your browser at http://localhost:8090/#/chat (pass --no-open to skip).

Next: the Quickstart.