Ollama
Ollama lets you run open-source models locally on your own hardware. No API key needed, no data leaves your machine.
Prerequisites
Section titled “Prerequisites”- Install Ollama - Download from ollama.com
- Pull a model - Run
ollama pullto download a model:
ollama pull llama3.1- Start Ollama - Make sure the Ollama server is running (it starts automatically on install)
Setup in Multi
Section titled “Setup in Multi”- Open the Multi panel → Settings (gear icon)
- Click Add Profile
- Select Ollama as the provider
- Enter the model name (e.g.,
llama3.1) - Save
The default base URL is http://localhost:11434. Change it if your Ollama server runs on a different address.
Configuration Options
Section titled “Configuration Options”| Option | Description |
|---|---|
| Base URL | Ollama server URL (default: http://localhost:11434) |
| Model ID | The model name (e.g., llama3.1) |
- GPU recommended - Models run much faster with a GPU
- RAM requirements - 7B models need ~8GB RAM, 70B models need ~40GB
- No internet needed - Once a model is downloaded, everything runs locally
- Privacy - No data leaves your machine
Free - Ollama is open-source and runs on your own hardware.