Skip to content

Open WebUI

Open WebUI is a user-friendly chat interface for Large Language Models (LLMs). It is compatible with Ollama, llama-swap, and the OpenAI API.

This is ONLY the Web Interface

Open WebUI is just a user interface. You additionally need an LLM server like Ollama or llama-swap to load models and process requests.

Installation

Add the following template to your docker-compose.yml and then run ei23 dc.

LLM Server Required

First install either Ollama or llama-swap.

Template

  open-webui:
    image: ghcr.io/open-webui/open-webui:main
    container_name: open-webui
    restart: always
    ports:
      - 2280:8080
    environment:
      - OLLAMA_BASE_URL=http://llama-swap:8080  # or http://ollama:11434
    volumes:
      - ./volumes/open-webui:/app/backend/data

Configure LLM Server

Depending on which LLM server you use, adjust the OLLAMA_BASE_URL:

LLM Server OLLAMA_BASE_URL Port in Template
llama-swap (recommended) http://llama-swap:8080 9292
Ollama http://ollama:11434 11434

llama-swap Recommended

For better performance and more flexibility, we recommend llama-swap over Ollama.

Features

  • Chat Interface - Modern, intuitive user interface
  • Multi-User Support - Multiple users with individual settings
  • Chat History - Conversations are saved and searchable
  • Prompt Templates - Reusable system prompts
  • RAG (Retrieval Augmented Generation) - Upload documents and ask questions about them
  • Model Selection - Switch between different models
  • OpenAI Compatible - Works with all OpenAI API-compatible servers

First Start

  1. After startup, you can access Open WebUI at http://[IP]:2280
  2. Create an account (the first account automatically becomes admin)
  3. Go to Settings → Connections and check the LLM server URL
  4. Select a model from the dropdown menu and start chatting

Models Must Be Downloaded First

Models are downloaded in the respective LLM server (Ollama/llama-swap), not in Open WebUI.

Upload Documents (RAG)

Open WebUI supports uploading documents to answer questions with context from your files:

  1. Click the 📎 icon in the chat
  2. Upload PDFs, text files, or Markdown files
  3. Ask questions about the content - the AI uses the documents as context

Notes

  • Data is stored in ./volumes/open-webui/
  • Port 2280 is configured by default - adjust if needed
  • Open WebUI also works with remote servers (cloud GPUs, etc.)
  • For best RAG results: Use a model with a large context length

Further Information