LlamaBarn 🦙 🌾

Run local LLMs on your Mac with a simple menu bar app. Launch any model with a single click, then chat with it via the built-in web UI or connect to it via the built-in REST API. LlamaBarn automatically configures models based on your Mac’s hardware to ensure optimal performance and stability.

Install with brew install --cask llamabarn or download from Releases ↗

Goals

• Make it easy for everyone to use local LLMs. Using local LLMs should not require technical knowledge. You should be able to just select a model from a list and start using it. Technical customizations should be possible, but not required.
• Make it easy for developers to add support for local LLMs to their apps. Adding support for local LLMs should be just as easy as adding support for cloud-based LLMs. You shouldn’t have to implement custom UIs for managing models, starting servers, etc.

Features

• Tiny (~12 MB) macOS app built in Swift
• Curated model catalog
• Automatic model configuration based on your Mac’s hardware
• Simple web UI that lets you chat with the running models
• Familiar REST API that lets you use the running models from other apps
• No side effects – installed models live in ~/.llamabarn and nothing is installed system-wide

Quick start

To get started:

• Click on the menu bar icon to open the menu
• Select a model from the catalog to install it
• Select an installed model to run it — the app will figure out the optimal model settings for your Mac and start a local server at http://localhost:2276

Use the running models in two ways:

• In the browser via the built‑in web UI
• In other apps via the REST API

API endpoints

LlamaBarn builds on the llama.cpp server (llama-server) and supports the same API endpoints:

# check server health
curl http://localhost:2276/v1/health

# chat with the running model
curl http://localhost:2276/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{"messages": [{"role": "user", "content": "Hi"}]}'

Find the complete reference in the llama-server docs ↗

Roadmap

• Embedding models
• Completion models
• Run multiple models at once
• Parallel requests
• Vision for models that support it