Ollama

Getting Started

1. Install Ollama from ollama.com with the one-line installer for macOS, Linux, or Windows.
2. Run your first model with a single command: ollama run llama3.2 to download and start chatting.
3. Browse available models at ollama.com/library and pull any model with ollama pull model-name.
4. Use the OpenAI-compatible API at localhost:11434 to integrate local models into your applications.

Key Features

• One-command setup downloads and runs any supported model with a single ollama run command.
• Extensive model library hosts hundreds of models including Llama, Mistral, Gemma, Phi, Qwen, and DeepSeek.
• OpenAI-compatible API serves models locally with an API that works as a drop-in replacement for OpenAI.
• Automatic quantization optimizes models for your hardware, running efficiently on consumer GPUs and Apple Silicon.
• Modelfile customization lets you create custom model configurations with system prompts, parameters, and templates.
• Cross-platform support runs natively on macOS (Apple Silicon), Linux (NVIDIA/AMD), and Windows.