Open WebUI
Self-hosted chat interface for local models via Ollama and cloud models via API, with knowledge bases and web search.
What it does
Open WebUI is a chat interface you run on your own hardware. It connects to local models via Ollama and to any OpenAI-compatible API, which includes cloud providers. You build knowledge bases from your own documents (RAG), use web search, voice input and image analysis, and extend the system with Python plugins. You manage users and roles yourself. Installation works via Docker, pip, Kubernetes or as a desktop app for macOS, Windows and Linux.
Who it suits
Teams that want an internal AI chat without handing data to an outside platform, and anyone who wants a convenient way to use local models. Someone has to set up and maintain the server.
Data protection for businesses
Your data stays wherever you run Open WebUI, for example on your own servers in the EU or fully offline. According to the documentation, the software sends nothing to the vendor by default; only when you connect external models do prompts and responses go to that model provider, under its terms. Since no data reaches Open WebUI, it does not train on your input either. The vendor does not mention a data processing agreement; if needed, you sign one with the connected model provider. The enterprise license adds single sign-on, audit logs and support with an SLA, among other things.
Limits
With more than 50 users in 30 days, the Open WebUI branding must remain visible; for this reason the license has not been OSI-certified since version 0.6.6. Enterprise pricing is available on request only. Answer quality depends on the model and, for local models, on your hardware.