Guides & Resources

Complete guide to running AI models on iPhone. Learn which models work best, RAM limitations, speed expectations, and step-by-step Ollama setup.

Comprehensive guide covering MacBook Air vs Pro, M1-M5 chip comparison, and which AI models to run on your Apple Silicon Mac.

The model-size-to-memory matrix: ~0.6 GB per billion parameters at Q4. What each RAM tier from 8GB to 128GB actually runs, with Q4/Q8 sizes and the top pick per tier.

Quick start guide for installing Ollama, selecting models, and running your first local LLM on macOS.

Learn why offline AI matters for privacy, how to set up completely private AI with no internet required.

The local coding tier list: Qwen3.6 27B, Devstral Small 2 and more with sourced benchmark scores, plus continue.dev, aider and Zed setup.

FLUX, SDXL and SD 3.5 on Apple Silicon: which image models fit 8-64GB RAM, with HuggingFace-verified file sizes and the best apps.

How Ollama picks Metal, CUDA, ROCm or CPU, what happens when a model does not fit in VRAM, and how to check the real split with ollama ps.

What stays on your device with local inference, what still touches the network, and how to make a local AI setup fully airtight.

Yes, with the right model. Engine-derived picks that fit an 8GB memory budget, what an 8GB Mac cannot run, and when 16GB is worth it.