Skip to content

Open WebUI

What it is

Open WebUI is a user-friendly WebUI for Large Language Models (LLMs), designed to provide a feature-rich, self-hosted chat interface.

What problem it solves

It provides a polished, ChatGPT-like interface for local LLMs (via Ollama) and external APIs, making them accessible to non-technical users. It adds features like RAG, multi-user support, and image generation that basic CLIs lack.

Where it fits in the stack

User Interface / Frontend. It sits on top of inference engines like Ollama or LiteLLM to provide the chat experience.

Typical use cases

  • Self-Hosted AI Chat: A private alternative to ChatGPT for family or organization use.
  • Local RAG: Uploading documents to chat with them using local embeddings and LLMs.
  • Model Comparison: Chatting with multiple models side-by-side to compare performance.

Strengths

  • Beautiful UI: Modern, responsive, and customizable.
  • Local RAG Support: Built-in support for document ingestion and retrieval.
  • Role-Based Access Control: Multi-user support with admin controls.
  • Wide Compatibility: Works with Ollama, OpenAI-compatible APIs, and more.
  • Channels & Streaming: (v0.9.0+) Support for real-time model streaming in Channels with full tool and RAG support.

Limitations

  • Resource Heavy: Requires its own resources alongside the inference engine.
  • Setup Complexity: RAG and advanced features require additional configuration (embeddings, vector stores).

When to use it

  • When you want a professional chat interface for your local models.
  • If you need to share access to your LLM server with other people securely.
  • For local document-based question answering (RAG).

When not to use it

  • If you prefer a minimal CLI-only workflow.
  • If you have extremely limited system resources.

Licensing and cost

  • Open Source: Yes (MIT License)
  • Cost: Free
  • Self-hostable: Yes

Getting started

Installation with Ollama (Docker Compose)

This example shows how to run Open WebUI and link it to an Ollama instance.

services:
  ollama:
    image: ollama/ollama:latest
    container_name: ollama
    volumes:
      - ./ollama:/root/.ollama
    restart: unless-stopped

  open-webui:
    image: ghcr.io/open-webui/open-webui:main # v0.9.0+ (May 2026)
    container_name: open-webui
    volumes:
      - ./open-webui:/app/backend/data
    depends_on:
      - ollama
    ports:
      - 3000:8080
    environment:
      - 'OLLAMA_BASE_URL=http://ollama:11434'
      # Security (v0.8.11+ / v0.9.0+)
      - 'AIOHTTP_CLIENT_ALLOW_REDIRECTS=false' # SSRF protection
      - 'IFRAME_CSP=default-src '\''self'\''; script-src '\''none'\'';' # Artifacts/Previews sandbox
    restart: unless-stopped

Security Posture (May 2026 Updates)

Open WebUI has introduced critical security hardening in recent versions (v0.8.11 and v0.9.0). - SSRF Protection: Redirects are blocked by default for all outbound HTTP requests (web fetch, image loading, tool execution) to prevent server-side request forgery. - Iframe Sandboxing: Administrators can now enforce a Content-Security-Policy for tool embeds and artifacts via the IFRAME_CSP environment variable. - CVE Fixes: v0.8.11 addressed an information disclosure flaw (CVE-2026-45666) in the notes API; v0.9.0 addressed an authorization bypass (CVE-2026-45671) in file management.

RAG & Knowledge Bases

Open WebUI provides a native "Knowledge" workspace to manage documents for Retrieval-Augmented Generation (RAG).

Setting Up a Knowledge Base

  1. Navigate to Workspace > Knowledge.
  2. Click the + icon to create a new Knowledge Base.
  3. Upload files (PDF, CSV, TXT, etc.) or sync an entire directory.
  4. Focused Retrieval: Open WebUI will chunk and embed your documents using a local embedding model (default: sentence-transformers/all-MiniLM-L6-v2) and store them in its internal ChromaDB.

Using Knowledge in Chat

  • Direct Mention: Type # in the chat bar to select a specific file or an entire Knowledge Base to include in the context.
  • Model Scoping: Go to Workspace > Models, edit a model, and attach a Knowledge Base. This model will now automatically use that knowledge for every query.

External Vector Databases

For large scale RAG, Open WebUI supports external vector stores like Milvus or Qdrant via environment variables: 

environment:
  - RAG_EMBEDDING_ENGINE=ollama
  - RAG_EMBEDDING_MODEL=mxbai-embed-large
  - VECTOR_DB=milvus
  - MILVUS_URI=http://milvus:19530

Backlog

  • [x] Perform quarterly technical freshness audit (May 2026).

Sources / References

Contribution Metadata

  • Last reviewed: 2026-05-26
  • Confidence: high