# Ollama Setup

Open Comet supports Ollama via the OpenAI-compatible API at `http://127.0.0.1:11434/v1/`

1. **Install Ollama**  
   Download and install Ollama from ollama.com. Available for macOS, Linux, and Windows.

2. **Start the server**  
   `ollama serve`

3. **Pull a model**  
   For text tasks use `llama3.2:3b`. For full browser-agent vision use `qwen2.5vl:7b`.
   
   `ollama pull qwen2.5vl:7b`

4. **Configure Open Comet**  
   In the Open Comet side panel: Settings → Provider → Ollama → Base URL: `http://127.0.0.1:11434`

5. **Refresh model list**  
   Click "Refresh models" to see your downloaded models. Select text and vision models (or one multimodal for both).