You can run AI models on your own computer, connect desktop apps, and use coding agents following official guidance from Ollama.
What it does
According to Ollama, you can use Ollama in desktop apps and coding agents, or build an application with an API key. You can also run models locally without an API key.
Before you start
Ollama states that the Gemma 4 E2B model download is about 7.2 GB. Ollama recommends 8 GB of available VRAM, or unified memory on a Mac, noting that larger context windows need more memory and smaller VRAM may result in slower responses.
Steps
Step 1: Download Ollama for macOS, Windows, or Linux, and open the app or get started from your terminal.
Step 2: To run a model locally, Ollama instructs you to run ollama run gemma4:e2b to download the model and start a chat.
What the company says
Ollama notes that cloud requests do not require an Ollama installation, and you should keep your API key on your application server outside browser code and source control.
Key facts and where they come from
- The Gemma 4 E2B model download is about 7.2 GB.
The model download is about 7.2 GB.
- Ollama recommends 8 GB of available VRAM or unified memory on a Mac.
We recommend 8 GB of available VRAM, or unified memory on a Mac.
