The Self-Hosted ChatGPT You Control

Open WebUI is a polished, self-hosted interface for chatting with large language models—without locking you into a single vendor. It works with Ollama and Llama.cpp for fully local, offline inference and with any OpenAI-compatible API, so you can mix local models with cloud providers such as OpenRouter, Groq, vLLM, or Mistral.

The project has become one of the most popular open-source AI UIs, with well over 150,000 GitHub stars. Setup is straightforward: a single Docker command, pip install open-webui, Kubernetes, or a desktop app. Once running, you get a ChatGPT-style experience plus extras that matter in real use: file uploads and RAG over your own documents, web search, tool calling, plugins, image generation, and granular user roles for teams.

That combination—privacy by default, provider flexibility, and a feature-rich UI—is the point. You can keep data on your machine, swap models without rewriting workflows, and extend the platform instead of fighting a closed chat box.

If you want a capable front door for local LLMs or a unified console for many APIs, start at github.com/open-webui/open-webui and the official docs at docs.openwebui.com.