Deploy and Explore Open WebUI: A Feature‑Rich Self‑Hosted AI Platform

What Is Open WebUI?

Open WebUI is an extensible, user‑friendly, self‑hosted AI platform designed for completely offline operation. It supports a variety of LLM runtimes—including Ollama and any OpenAI‑compatible API—and includes a built‑in reasoning engine for Retrieval‑Augmented Generation (RAG), making it a powerful solution for deploying AI services.

Main Features

Easy setup via Docker or Kubernetes (kubectl, kustomize, or Helm) with support for :ollama and :cuda tagged images.

Ollama/OpenAI API integration: connect to OpenAI‑compatible endpoints, LMStudio, GroqCloud, Mistral, OpenRouter, etc.

Granular permissions and user groups for secure, customized user experiences.

Responsive design for desktop, laptop, and mobile devices.

Progressive Web App (PWA) support for offline, native‑like mobile usage.

Full Markdown and LaTeX rendering for rich LLM interactions.

Voice/video chat using multiple speech‑to‑text (local Whisper, OpenAI, Deepgram, Azure) and text‑to‑speech providers (Azure, ElevenLabs, OpenAI, Transformers, WebAPI).

Model builder: create Ollama models, add custom roles/agents, and import community models.

Local Python function calls (BYOF) via an integrated code editor.

Persistent artifact storage with a key‑value API for logs, trackers, leaderboards, and shared data.

Local RAG integration with nine vector‑database options and content extractors (Tika, Docling, Mistral OCR, etc.).

Web search for RAG using over 15 providers (SearXNG, Google PSE, Brave Search, Kagi, Mojeek, Perplexity, Serpstack, Serply, DuckDuckGo, SearchApi, SerpApi, Bing, Jina, Exa, Sougou, Azure AI Search, Ollama Cloud).

Web browsing via # commands to inject URLs directly into chats.

Image generation and editing with DALL‑E, Gemini, local ComfyUI, and AUTOMATIC1111.

Multi‑model conversations to leverage the strengths of different LLMs simultaneously.

Role‑Based Access Control (RBAC) for secure model management.

Flexible database/storage options: SQLite (optional encryption), PostgreSQL, or cloud back‑ends (S3, GCS, Azure Blob).

Advanced vector‑database support (ChromaDB, PGVector, Qdrant, Milvus, Elasticsearch, OpenSearch, Pinecone, S3Vector, Oracle 23ai).

Enterprise authentication: LDAP/AD, SCIM 2.0, SSO via trusted headers, OAuth providers (Okta, Azure AD, Google Workspace).

Cloud‑native file pickers for Google Drive and OneDrive/SharePoint.

Production observability with built‑in OpenTelemetry (traces, metrics, logs).

Horizontal scalability using Redis‑based session management and WebSocket support for multi‑worker/node deployments.

Internationalization (i18n) for multiple languages and a call for contributors.

Pipeline and plugin framework to integrate custom Python logic, rate limiting, Langfuse monitoring, LibreTranslate, malicious‑message filtering, etc.

Regular updates with bug fixes and new features.

Installation Guides

Python pip Installation

Open WebUI can be installed with pip. Ensure you are using Python 3.11 to avoid compatibility issues.

Install the package: pip install open-webui Start the server: open-webui serve Access the UI at http://localhost:8080.

Docker Quick Start

For some Docker environments additional configuration may be required; refer to the official Open WebUI documentation for troubleshooting.

When using Docker, mount the data volume to preserve your database: -v open-webui:/app/backend/data

To enable CUDA acceleration, use the :cuda image tag and install the NVIDIA Container Toolkit on your Linux/WSL host.

Default Docker Configuration

If Ollama runs on the same host:

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

If Ollama is on a remote server, set the environment variable OLLAMA_BASE_URL to the remote URL:

docker run -d -p 3000:8080 -e OLLAMA_BASE_URL=https://example.com -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main

For GPU‑enabled containers:

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

OpenAI‑Only Docker Setup

When using only the OpenAI API, provide your secret key via the OPENAI_API_KEY environment variable:

docker run -d -p 3000:8080 -e OPENAI_API_KEY=your_secret_key -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main

Bundled Ollama Docker Image

This image combines Open WebUI and Ollama in a single container for simplified deployment.

GPU‑enabled version:

docker run -d -p 3000:8080 --gpus=all -v ollama:/root/.ollama -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:ollama

CPU‑only version:

docker run -d -p 3000:8080 -v ollama:/root/.ollama -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:ollama