OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Infrastructure & Serving

Mlx Serve

Mlx Serve is a self-hosted language-model inference server for Apple Silicon, implemented in Zig and supporting MLX and GGUF models. It provides OpenAI-, Anthropic- and Ollama-compatible local APIs, alongside a macOS app and command-line interface.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Mlx Serve — OpenViST project profile in AI Infrastructure & Serving.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Mlx Serve when you want to host MLX or GGUF model inference locally on Apple Silicon and connect software that speaks one of its supported API formats.

Who it is for

  • Apple Silicon users who want to run language models locally.
  • Developers using clients or SDKs that support OpenAI, Anthropic or Ollama-compatible APIs.
  • Users who want a macOS app for managing models and chatting with a local server.

Documented uses

  • Serve MLX or GGUF language models from a local machine.
  • Connect compatible clients to a local inference endpoint.
  • Download and manage models through the macOS app or CLI.
  • Build the app and server from source using the documented macOS toolchain.
Before you choose

Requirements & limits

The documented Apple Silicon app path requires macOS 26.2 or later. The repository also provides Homebrew installation and separate Linux NVIDIA release builds.

  • The Apple Silicon app and server require macOS 26.2 or later; Linux NVIDIA builds have a distinct driver, CUDA and library prerequisite list.
  • The macOS app bundles the server and provides model browsing, downloads, chat, agent mode with MCP tool calling, and server settings.
  • The CLI supports model download, listing and serving; the repository documents source builds requiring Xcode 26.2 or later and Homebrew.
  • The local endpoint is documented as http://localhost:11234, with OpenAI-, Anthropic- and Ollama-compatible APIs.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install MLX-Serve with Homebrew

Install the MLX-Serve macOS app or its CLI and server.

Before you start

  • macOS 26.2 or later on Apple Silicon
  • Homebrew

Steps

  1. Add the tap: brew tap ddalcu/mlx-serve https://github.com/ddalcu/mlx-serve.

  2. Trust the tap: brew trust ddalcu/mlx-serve.

  3. Install the app with brew install --cask mlx-serve, or install the CLI and server only with brew install mlx-serve.

  4. To upgrade, run brew update followed by brew upgrade --cask mlx-serve for the app or brew upgrade mlx-serve for the CLI.

Expected outcome

The selected app or CLI/server package is installed; the app includes the server.

Guide limitations

  • The documented prerequisites specify macOS 26.2 or later on Apple Silicon.
  • The app and CLI are separate package choices.
Documentation 1 ↗
Source-based practical guide

Build MLX-Serve app and server from source

Build the macOS app and server from the repository source.

Before you start

  • macOS with Xcode 26.2 or later
  • Homebrew

Steps

  1. Clone the repository and its submodules: the command below.

    Documented command
    git clone --recurse-submodules https://github.com/ddalcu/mlx-serve && cd mlx-serve

  2. Run ./app/build.sh to build the app and server.

Expected outcome

The script builds the app and server with ad-hoc signing.

Guide limitations

  • The documented build target is macOS.
  • The script may download the Metal Toolchain component and install the Brewfile dependencies, cmake and webp, when missing.
  • Server-only build instructions are referenced in the repository's building documentation.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Model inference and serving

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
Mlx ServeServe MLX or GGUF language models from a local machine.macOS, Self-hostedThe documented Apple Silicon app path requires macOS 26.2 or later. The repository also provides Homebrew installation and separate Linux NVIDIA release builds.
MorphingdbImport a libtorch model and create predictions from values in a PostgreSQL table.Docker, Linux, Self-hosted, ServerThe README documents a Docker image and a build-from-repository option, followed by connecting to PostgreSQL and enabling the pgdl extension.
Gateway — pyadavSend chat-completion requests to a provider through a common server endpoint.ServerThe repository is public and identifies the project as Apache-2.0 licensed. README installation instructions include binaries, platform package managers and Docker.
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
14 October 2026
Original website title
ddalcu/mlx-serve · Official homepage title · retained in its source language
How we review sourcesSuggest a correction