Run models locally

Ollama

Download and chat with open AI models on your own PC with one command, no cloud account needed.

Rank

#2

Stars

186k

Licence

MIT

Checked

6 Nov 2026

What it does

Ollama is a small app that downloads open-weight language models (Llama, Gemma, Mistral, Qwen and others) and runs them on your own computer. It gives you a chat in the terminal and a local API on port 11434, so other apps can use the model as if it were a cloud service. Nothing you type leaves your machine.

How to use it

Shell: PowerShell (Windows 11)

1. 1

   Install it with winget, or download the Windows installer from ollama.com.

   winget install Ollama.Ollama

2. 2

   Download a small model and start chatting. Type /bye to quit.

   ollama run llama3.2

3. 3

   Call the same model from any script through the local API.

   Invoke-RestMethod http://localhost:11434/api/generate -Method Post -Body '{"model":"llama3.2","prompt":"Explain DNS in one line","stream":false}' | Select-Object -ExpandProperty response

4. 4

   List and remove models to free disk space.

   ollama list
   ollama rm llama3.2

Best practices

• Pick a model that fits your memory. A 3B model runs on most laptops; 7-8B models want about 8 GB of free RAM or VRAM.
• Keep the API on localhost. Never open port 11434 to the internet, because it has no login by default.
• Pin a model tag (for example llama3.2:3b) in scripts so an update doesn't silently change your results.
• Check each model's licence on its library page before using it in a commercial product.

In the news

• TechCrunch · 9 Jul 2026 · Popular open source AI developer tool Ollama raises $65M, grows to nearly 9M users
• The New Stack · 31 Mar 2026 · Ollama taps Apple's MLX framework to make local AI models faster on Macs