Ollama

Open-source model server for running large language models locally on your own hardware. Install with a single command and run Gemma, Qwen, DeepSeek, GLM, and dozens of other open models without sending data to the cloud.

Ollama runs natively on macOS, Windows, and Linux. It exposes a REST API on localhost and supports Python and JavaScript SDKs, plus integrations with LangChain, LlamaIndex, n8n, Dify, and Open WebUI. Unlike cloud-hosted APIs, models run entirely on your machine for full privacy and offline use.

Key capabilities:

• One-line install and pull models via CLI
• REST API and official Python and JavaScript libraries
• Interactive menu to launch coding tools such as Claude Code and OpenClaw
• Works with LangChain, LlamaIndex, n8n, and other frameworks
• Runs on CPU or GPU, supports NVIDIA, AMD, and Apple Silicon

Engineers use Ollama for local prototyping and development, embedding models into apps via the API, building RAG and document workflows with LangChain or LlamaIndex, and running coding assistants or automation flows without cloud dependencies.