Open WebUI
A self-hosted chat interface that works with Ollama and any OpenAI-compatible API, built to run entirely offline.
Open WebUI is a web-based chat platform that you host yourself. It describes itself as extensible, easy to use and built to run entirely offline. You connect local models through Ollama, or any OpenAI-compatible API for local and cloud models, and then use it from a browser. It installs with pip or uvx, or runs as a Docker container; separate images exist with local AI dependencies, a slim version that relies on external services, and one bundled with Ollama.
It is a good fit when several people need a ChatGPT-style interface on top of models you control, for example a team or family sharing one machine or server.
The license needs a careful read. Open WebUI uses its own license, based on the BSD 3-Clause text, with an extra condition that you must not remove or alter the Open WebUI name and logo. The exceptions are deployments with no more than 50 end users in any 30 days, written permission from the project, or an enterprise license. Older contributions keep their original licenses. Because of that branding condition it is free to use but not open source in the strict sense.
Good for
- A shared, browser-based chat interface for a team or household
- Putting a friendly front end on Ollama or any OpenAI-compatible endpoint
- Setups that must run on your own servers and offline
Watch out for
- The branding clause matters if you rebrand it or serve more than 50 users in 30 days; read the license first
- It is a front end only, so you still need a model server such as Ollama
- Running it for many users means you take on updates, backups and access control