Start here · Step 1 of 7

1. What is a local LLM
2. Hardware
3. Words
4. Pick a model
5. Install
6. First chat
7. Next

Step 1 of 7 · about 5 minutes

What is a local LLM

A language model is a program that predicts the next word. “Local” means the weights live on a machine you have, and the prompt does not have to leave the house.

What you gain: privacy, no per-token bill, a machine that still works when the network is down, and the chance to measure it yourself.

What you give up: the biggest cloud models, and speed that depends on the hardware in front of you. A laptop and a used server will not feel the same. I will not tell you they will.

This site’s job is to write down what my machines actually do, and to leave a cell blank when I have not measured it. Other people’s numbers, when they appear, stay labeled as theirs.