Llama

Llama is a macOS menu bar app for running local LLMs.

Watch a 2-minute intro 📽️

Install

brew install --cask llama-app

Or download from Releases.

How it works

Llama runs a local server at http://localhost:8080/v1.

• Install models — from the recommended list or Hugging Face
• Connect any app — chat UIs, editors, CLI tools, scripts
• Models load when requested — and unload when idle

Features

• 100% local — Models run on your device; no data leaves your Mac
• Small footprint — 4 MB native macOS app
• Zero configuration — models are auto-configured with optimal settings for your Mac
• Recommended for your Mac — suggests models sized to fit your hardware; browse the full catalog at llama.app
• Standard storage — models live in the Hugging Face cache (~/.cache/huggingface/hub), shared with llama.cpp and other HF-aware tools
• Built on llama.cpp — from the GGML org, developed alongside llama.cpp

Works with

Llama works with any OpenAI-compatible client.

• Chat UIs — Chatbox, Open WebUI, BoltAI (instructions)
• Editors — VS Code, Zed, Xcode (instructions)
• Editor extensions — Cline, Continue
• CLI tools — OpenCode (instructions), Claude Code (instructions)
• Custom scripts — curl, AI SDK, etc.

API examples

# list installed models
curl http://localhost:8080/v1/models

# chat with Gemma 3 4B (assuming it's installed)
curl http://localhost:8080/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model": "gemma-3-4b", "messages": [{"role": "user", "content": "Hello"}]}'

Replace gemma-3-4b with any model ID from http://localhost:8080/v1/models.

See complete API reference in the llama.cpp server docs.

Experimental settings

Expose to network — By default, the server is only accessible from your Mac (localhost). This option allows connections from other devices on your local network. Only enable this if you understand the security risks.

# bind to all interfaces (0.0.0.0)
defaults write app.llama.Llama exposeToNetwork -bool YES
or bind to a specific IP (e.g., for Tailscale)
defaults write app.llama.Llama exposeToNetwork -string "100.x.x.x"
disable (default)
defaults delete app.llama.Llama exposeToNetwork

Roadmap

• [ ] Support for loading multiple models at the same time
• [ ] Support for multiple configurations per model (e.g., multiple context lengths)