Guides

Step by step: local or API

Choose between running models on your own machine (private, offline, free) or consuming via API from major providers.

Local LLM

Runs on your hardware. Zero cost per token, data never leaves.

Ollama · Beginner
Run LLMs locally on macOS with Ollama

Install Ollama, download a model, and make your first local call in under 10 minutes without sending data anywhere.

8 min· macOS Apple Silicon (M1/M2/M3/M4)
LM Studio · Beginner
LM Studio on Windows - a graphical way to run local LLMs

If you prefer clicking over terminal commands, LM Studio is the fastest path on Windows.

7 min· Windows 10/11 x64
llama.cpp · Advanced
llama.cpp on Linux - maximum performance without overhead

Building from source gives you the highest throughput and fine-grained GPU flags.

10 min· Ubuntu 22.04+
Provider APIs

Set up OpenAI, Anthropic, Groq and others in minutes.

OpenAI · Beginner
OpenAI API - first call, streaming, and cost control

Go from zero to a production feature with GPT-5 without blowing up the bill.

9 min
Anthropic · Beginner
Anthropic Claude API - from the first token to tool use

Claude 4.5 Sonnet is a strong default for coding and agents. Here's how to connect it.

8 min
Groq · Beginner
Groq API - extremely fast inference with a free tier

Llama, Qwen, and Mixtral at 500+ tokens per second. Perfect for chatbots with instant UX.

6 min