100% Private — Runs Offline

AI That Lives on Your Device

Run powerful language models locally on your Mac with Metal GPU acceleration. No cloud. No subscriptions. No data leaves your machine.

macOS 13+ • Apple Silicon & Intelv2.0.0 • ~2.5 GB
ALI Edge — Chat

Why ALI Edge

Everything runs locally. Your data never touches a server.

Metal GPU Acceleration

Leverages Apple’s Metal framework for blazing-fast inference on M-series and Intel Macs.

Completely Private

Zero telemetry. Zero cloud calls. Your conversations stay on your hardware — always.

7 Expert Models

Specialized experts for code, math, writing, analysis, and more — routed automatically.

Zero Configuration

Download, double-click, done. Python, models, and dependencies are all bundled in.

System Dashboard

Monitor CPU, RAM, disk, and GPU usage in real-time while your models are running.

Dark & Light Themes

Beautiful neumorphic UI with both dark and light modes. Feels native on macOS.

Under the Hood

Built with performance-critical Rust orchestration and Python AI serving.

Inference Engine
llama.cpp + Metal
Router Model
Qwen 2.5 · 0.5B
Expert Model
Qwen 2.5 · 3B Instruct
Orchestration
Rust (arm64 + x86)
API Server
FastAPI + Uvicorn
Format
GGUF Quantized
Min macOS
13.0 Ventura
License
MIT

Get Started in 3 Steps

1

Download

Grab the DMG — everything is bundled: Python, models, dependencies.

2

Install

Drag ALI Edge to your Applications folder. That’s the entire install.

3

Chat

Double-click to launch. Your private AI assistant is ready in seconds.

Download ALI Edge

Version 2.0.0 • macOS 13+

Download DMG
Size~2.5 GB
ArchitectureApple Silicon & Intel
IncludesPython, Models, Runtime

No account required. No internet needed after install.