Install Open WebUI (Linux) – One-Liner Installer Script with Docker & Caddy HTTPS

Try Open WebUI first and then run it in your organisation?

The script gets Open WebUI running in minutes; updates, backups and monitoring stay with you afterwards. If you would rather hand that over, we run Open WebUI as the AI platform: a dedicated instance with models operated in Germany by German providers, SSO and a Section 203 supplementary agreement, from €149.90 excl. VAT per month. For data that must not leave the building, the AI Cube brings Open WebUI preconfigured into your own network.

View the AI platform · Schedule an initial consultation

Open WebUI is a powerful self-hosted web interface for Large Language Models. Whether you want to run local models with Ollama or connect to OpenAI, Open WebUI gives you a ChatGPT-like experience on your own infrastructure.

We've created a free, open-source installer script that deploys Open WebUI with Docker in minutes - complete with optional HTTPS, GPU support, and production-ready defaults.

Table of Contents

• What You Get
• Prerequisites
• Option 1: Quick Install (One Command)
• Option 2: Review Before Installing (Recommended)
• Installation Walkthrough
• Post-Installation Setup
• Useful Commands
• Connecting to LLM Backends
• Uninstallation
• Troubleshooting
• Source Code

What You Get

Our installer script provides:

• Automatic Docker installation - Docker and Docker Compose are installed if not present
• Interactive configuration - Choose your installation directory, domain, GPU support, and image variant
• Optional HTTPS - Caddy reverse proxy with automatic Let's Encrypt certificates
• GPU support - NVIDIA GPU acceleration for local model inference
• Production-ready setup - Sensible defaults for immediate use

Prerequisites

Before running the installer, ensure you have:

• A Linux server (Ubuntu, Debian, CentOS, RHEL, etc.)
• bash and curl installed (standard on most systems)
• sudo access for Docker installation
• For HTTPS: A domain name with DNS pointing to your server
• For GPU: NVIDIA GPU with nvidia-docker installed

Option 1: Quick Install (One Command)

If you trust the source, deploy Open WebUI with a single command:

curl -fsSL https://raw.githubusercontent.com/wzitcom/public-scripts/main/open-source-apps/open-webui/install.sh | bash

Option 2: Review Before Installing (Recommended)

For security-conscious users, download and review the script first:

# Download the installer
curl -O https://raw.githubusercontent.com/wzitcom/public-scripts/main/open-source-apps/open-webui/install.sh

# Review the script contents
cat install.sh

# Make it executable and run
chmod +x install.sh
./install.sh

We always recommend reviewing scripts before execution, especially when running with elevated privileges.

Installation Walkthrough

When you run the installer, you'll see an interactive setup process:

The script guides you through:

1. Installation directory - Where to store configuration files (default: ~/open-webui)
2. Domain configuration - Optional domain for HTTPS setup
3. GPU support - Enable NVIDIA GPU acceleration
4. Image selection - Choose between full or slim Docker image

Note (as of October 2026): The slim variant only exists without CUDA. If you enable GPU support, pick the full image - combining GPU and slim leads to a pull error because the corresponding tag does not exist upstream.

After configuration, the script:

• Installs Docker if needed
• Creates docker-compose.yml with your settings
• Sets up Caddy for HTTPS (if domain provided)
• Starts the containers
• Displays access information

Post-Installation Setup

Once installation completes, you'll see the Open WebUI login screen:

First Login

1. Navigate to your Open WebUI instance (http://localhost:3000 or your domain)
2. Click "Sign up" to create an account
3. The first user automatically becomes Administrator

Installation Files

Your installation directory contains:

Useful Commands

Navigate to your installation directory first:

cd ~/open-webui

Service Management

# View logs
docker compose logs -f
docker compose logs -f open-webui
docker compose logs -f caddy  # If using Caddy

# Stop services
docker compose down

# Start services
docker compose up -d

# Restart services
docker compose restart

# Check status
docker compose ps

Updates

# Pull latest images and restart
docker compose pull
docker compose up -d

Connecting to LLM Backends

Ollama (Local Models)

If running Ollama on the same machine:

cd ~/open-webui
nano .env

Add or uncomment:

OLLAMA_BASE_URL=http://host.docker.internal:11434

Important on Linux (as of October 2026): Docker Desktop resolves host.docker.internal automatically, Docker Engine on Linux does not. The docker-compose.yml generated by the script contains no matching entry, so the name does not resolve. Add the following to the open-webui service in docker-compose.yml:

    extra_hosts:
      - "host.docker.internal:host-gateway"

The special value host-gateway points to the internal IP of the host (Docker documentation). Alternatively, put the host IP directly into OLLAMA_BASE_URL, for example http://172.17.0.1:11434 for the default Docker bridge.

Ollama also has to listen on that address: by default it binds only to 127.0.0.1:11434; the bind address can be changed via OLLAMA_HOST, with systemd via systemctl edit ollama.service (Ollama FAQ). Afterwards, port 11434 should only be reachable from the Docker network via firewall, not from the internet.

Then restart:

docker compose down && docker compose up -d

OpenAI API

Edit .env and add your API key:

OPENAI_API_KEY=sk-your-api-key-here

Restart the container to apply changes.

Instead of a US provider, you can also connect OpenAI-compatible models operated in German data centres of German providers. Ready to use with gpt-oss-120b, Qwen and Ministral, this is available as the AI platform.

Uninstallation

To completely remove Open WebUI:

cd ~/open-webui

# Stop and remove containers
docker compose down

# Also remove volumes (WARNING: deletes all data!)
docker compose down -v

# Remove installation directory
cd ~
rm -rf ~/open-webui

Troubleshooting

Port Already in Use

If ports 80/443 are occupied (when using Caddy):

# Check what's using the ports
sudo lsof -i :80
sudo lsof -i :443

SSL Certificate Issues

# Check Caddy logs
docker compose logs caddy

# Verify DNS
nslookup your-domain.com

GPU Not Detected

# Verify nvidia-docker is working
docker run --rm --gpus all nvidia/cuda:12.6.0-base-ubuntu24.04 nvidia-smi

Container Won't Start

# Check container logs
docker compose logs open-webui

# Check container status
docker ps -a

Source Code

The installer script is open source and available on GitHub:

• Repository: github.com/wzitcom/public-scripts
• Open WebUI Installer: open-source-apps/open-webui

Feel free to review, fork, or contribute!

Disclaimer

This script is intended for quickly deploying development and testing environments.

For production deployments, please consider:

• Proper security hardening
• SSL/TLS certificate configuration
• Firewall and network security rules
• Regular backup strategies
• High availability setup
• Resource monitoring and alerting
• Custom environment configurations

Need a production-ready setup? Talk to WZ-IT about Open WebUI deployment, integration and operations.