← Back to iRun

Mac

Local AI for Mac, Done Right

iRun Studio is a native macOS powerhouse built for Apple Silicon. Fast on-device inference, Model Browser with Language/Vision/Logic tabs, autonomous tool calling, image generation, and zero per-token costs.

Download for macOS

Mac-only power features

  • Unified Memory Acceleration: High-speed Metal backend running 4B to 70B+ GGUF and MLX models
  • Model Browser: Browse, download, and switch between Language, Vision, and Logic models
  • Multi-Engine Inference: Fast, Thinking, and Pro modes with full streaming telemetry
  • Native Tool Calling: Integrated on-device web search, shell execution, and file inspection
  • 120Hz ProMotion Subagent Bar: Real-time task execution and autonomous workflows
  • CoreML & Safetensors Diffusion: On-device image creation without sending prompts to the cloud

Recommended for

Developers, researchers, writers, and privacy-conscious Mac users who want ChatGPT-level utility without sending every prompt to OpenAI. Optional Claude, Gemini, and OpenRouter connections when you need frontier cloud models.

Setup Guide vs Cloud Comparison Offline AI Architecture Private AI for iPhone