Open WebUI Download: A ChatGPT-Style Interface for Local AI Models

Running a model in the terminal with Ollama is easy, but chat history, file uploads or several people sharing the same model all call for a proper interface. Open WebUI fills that gap: it is an AI platform that opens in your browser and runs on your own computer or server. Its GitHub repository has roughly 157,000 stars as of October 2026.

What does Open WebUI do?

Open WebUI does not run models itself; it is a front end that connects to the engines that do. You can use Ollama locally and services offering an OpenAI-compatible API in the cloud, side by side on the same screen.

Key features

• Familiar chat screen: Chat history, Markdown and LaTeX rendering, and the option to ask several models the same question at once.
• Chat with documents (RAG): Upload PDFs and similar documents and ask questions about their contents.
• Web search: Adds search results to answers using providers such as SearXNG, Brave and DuckDuckGo.
• Image generation: Creates images by connecting to engines such as ComfyUI, AUTOMATIC1111 or OpenAI.
• Multi-user: Share it with a team or family, with role, group and permission management.
• Extensions: Expand its capabilities with tools, filters and MCP servers.
• Voice: Hands-free conversations with speech-to-text and text-to-speech engines.
• Multilingual interface: The interface is available in many languages.

Who is it for?

It suits Ollama users who want a comfortable interface instead of the terminal, small teams who want to use AI without their data leaving the company, and enthusiasts who want to try different models in one place.

How to install

With Python (pip): The developers recommend Python 3.11 or 3.12; Python 3.13 is not supported yet.

pip install open-webui
open-webui serve

The interface opens at http://localhost:8080.

With Docker: If Ollama is installed on the same computer, this is the command from the official docs:

docker run -d -p 3000:8080 --add-host=host.docker.internal:host-gateway -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main

In this case the interface is at http://localhost:3000. Do not remove the -v open-webui:/app/backend/data part of the command: your chats and settings are stored in that volume, and without it your data may be lost when you update. There are also ready-made images that bundle Ollama, and CUDA-enabled images for NVIDIA graphics cards.

First steps

The first time you open the address, you are asked to create an administrator account. The first account becomes the admin, and new sign-ups are then turned off automatically; to add other people, re-enable sign-ups from the admin panel. If Ollama is running on your computer, your installed models appear in the list automatically and you can start chatting right away.

Hardware requirements

Open WebUI itself is lightweight; the real load falls on the engine running the model. If you run models locally with Ollama, your hardware needs are those of Ollama and the model you choose. If you only connect to cloud APIs, an ordinary computer or a small server is enough.

License

Open WebUI is free and its source code is open. Its license is the “Open WebUI License”, a BSD-3 license with an added branding clause: deployments serving more than 50 users in a 30-day period cannot change the “Open WebUI” name and logo without permission or an enterprise license. For personal and small-team use, this clause imposes no restriction.