Open WebUI - Self-Hosted ChatGPT Interface

Open WebUI (formerly Ollama WebUI) is a self-hosted, ChatGPT-like frontend for local and remote LLMs. Connect Ollama, LiteLLM, or any OpenAI-compatible endpoint and get conversations, document RAG, multi-user access, and extensions—without depending on a single vendor chat product.

Key Features

• Familiar Chat UX: Models, threads, and a clean interface users already understand.
• Flexible Backends: Ollama, OpenAI, Azure, Groq, or your LiteLLM proxy.
• RAG: Upload documents and chat with your own knowledge base.
• Multi-User: Auth and shared or private workspaces for teams.
• Extensible: Tools, functions, and pipeline-style integrations.
• Easy Hosting: Official Docker images for quick VPS deploys.

Why Choose Open WebUI?

• You run Ollama or LiteLLM and need a proper UI, not only CLI.
• ChatGPT is fine until data residency or cost becomes an issue.
• Teams want shared access to local models with accounts.
• Pairs with ToolHunt entries like Ollama (Mac), LiteLLM, and Langfuse.

Docker Deployment

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

Or use Docker Compose and set OLLAMA_BASE_URL / OpenAI-compatible env vars for your backend. See project docs for GPU and multi-service layouts.

Getting Started

1. Deploy Open WebUI and create the first admin user.
2. Connect Ollama (same host or remote) or an API key backend.
3. Pull models and start a chat.
4. Optional: enable document upload for RAG.
5. Invite users and set permissions.

Full Setup Guide

openwebui.com · github.com/open-webui/open-webui. For local models, see Ollama Docker install on BitDoze.