Llamaport for MAC

Llamaport is a small Mac app for running and downloading local GGUF language models. It wraps llama-server from the llama.cpp project: it lists the models in your models folder, launches one with the full command shown before you press start, shows what it costs while it runs, and fetches new models from Hugging Face with downloads that can resume after a kill or a reboot. It reads GGUF headers directly, so the list shows the real quantisation, context length and mixture-of-experts structure rather than guesses from the file name.

Llamaport is made by Muhammad Kamran and is free and open source under the MIT license. Version 0.7.0 is the first release after a redesign, with a sidebar for Library, Discover, Downloads, Activity Monitor and Settings, and the new Discover screen for browsing Hugging Face's GGUF repositories. Two downloads are offered: one for Apple Silicon and a universal one for Apple Silicon and Intel Macs. The developer says the universal build has never been run on an Intel Mac.

Llamaport is a beta and is not signed by Apple, so macOS blocks the first launch; use Open Anyway in System Settings, Privacy & Security. It does not include llama.cpp: you need llama-server installed, for example with Homebrew (brew install llama.cpp). It runs one model at a time, has no chat window of its own (it opens llama.cpp's Web UI) and only reaches public Hugging Face repositories.

Key Features

Llamaport takes the retyping out of running local models with llama-server.

• Model library: Lists your GGUF models with their real quantisation, context length and size.
• Visible launch command: Shows the full llama-server command before you start a model.
• Presets: Default, Best speed and Model suggested set the settings nobody wants to hand-tune.
• Test model: Checks the process, port, health endpoint, model list and a real chat completion.
• Memory sizing: Works out weights and KV-cache needs against what the GPU can actually use.
• Resumable downloads: Four ranged connections, sha256 checks and resume after a kill or reboot.
• Discover: Browse Hugging Face GGUF repositories with a quantisation picked to fit your Mac.
• Activity Monitor: Shows every llama-server on the Mac with its memory and CPU use.

Llamaport 0.7.0 is the first release since the redesign and the first with Discover.

• Discover browses Hugging Face's GGUF repositories live, with trending, most downloaded and most liked lists, filters and search, and picks a quantisation that fits your GPU.
• A new sidebar layout with Library, Discover, Downloads, Activity Monitor and Settings, and three launch presets on each model screen.
• Activity Monitor lists every llama-server on the Mac with its memory and CPU.
• Appearance settings with five palettes, light or dark, and a translucent sidebar.
• Network commands and the llama-server binary now run off the main thread, and the app's web view has a Content Security Policy.