Guides
Guides to AI on Apple silicon
Straight answers for developers buying inference and for Mac owners deciding whether to run or share models. Every number is sourced.
How to run an LLM on a Mac
How to run an LLM on a Mac with Apple Silicon: the four main tools, how to pick a model size for your memory, and a full mlx-lm walkthrough.
7 min read →How much RAM do you need to run an LLM?
How much RAM you need to run an LLM: the parameters × bits ÷ 8 rule, KV cache and macOS headroom, plus measured peak memory on Apple Silicon Macs.
7 min read →What is MLX?
What is MLX: Apple's open-source array framework for Apple silicon, how unified memory works, what mlx-lm does, and how 4-bit quantization with groups works.
6 min read →Mac mini for AI: which configuration fits which model?
Mac mini for AI and local LLMs: which memory configurations fit which model sizes, how it compares with Mac Studio and MacBook, and how an idle mini can earn.
7 min read →How to make money with an idle Mac
How to make money with an idle computer like a Mac: which networks support Apple Silicon, what Smarty Pants pays today, safety and taxes.
5 min read →What does "OpenAI-compatible API" actually mean?
What an OpenAI-compatible API is, which endpoints share one request shape, what still differs between providers, and Python and curl examples.
6 min read →Decentralized vs. centralized AI inference: what's the real trade-off?
What decentralized and distributed AI inference mean, how they compare with centralized providers on trust, latency, capacity and price.
7 min read →LLM API pricing explained, with a worked example
How LLM API pricing works: tokens, input vs. output price per million, reasoning tokens, embeddings and per-second compute, with a worked cost example.
7 min read →What is an AI agent sandbox, and how do code sandbox APIs work?
What an AI agent sandbox is, how code sandbox APIs handle commands, files, snapshots and network access, and how Smarty Pants CPU jobs and workspaces work.
7 min read →How can an AI agent find, try and pay for an AI API on its own?
How an AI agent finds an AI API, checks models and prices, stays inside a spending limit and gets a person to pay.
5 min read →How do I use open models in my agent framework?
Copy-paste setups for exact open models in LangChain, OpenAI Agents SDK, Vercel AI SDK, LlamaIndex and Claude Code, with listed prices.
5 min read →Use AI on real Macs, or let your Mac earn.
Developers call pinned builds of open models through an OpenAI-compatible API. Mac owners make an idle Apple silicon Mac available and get paid for the hours it is ready.
Requests run on Macs owned by other people. Owners agree not to read requests and the app doesn't show them, but a serving Mac decrypts each request in memory to answer it, so that is a rule and a design, not a technical guarantee. Don't send sensitive data on the general network.