How to Run Ollama with Open WebUI — Complete Setup Guide (2026)

The Ollama terminal is perfectly functional, but let’s be honest — typing into a command-line chat interface isn’t how most people want to use AI on a daily basis. Open WebUI changes that completely. It gives you a full ChatGPT-style browser interface connected directly to your local Ollama installation — private, offline, and completely free.

I’ve been running Open WebUI as my daily AI interface for over a year. In this guide, I’ll show you every installation method, the complete setup process, how to configure it properly, and the features that make it genuinely better than ChatGPT for many use cases.

What is Open WebUI?

Open WebUI (formerly known as Ollama WebUI) is a self-hosted, feature-rich web interface designed to work with Ollama and other local AI backends. Think of it as a private, offline version of ChatGPT that runs entirely on your own computer.

What you get with Open WebUI

• ChatGPT-style conversation interface — multi-turn conversations with clean formatting
• Model switcher — switch between any Ollama model from a dropdown
• Conversation history — all chats saved locally, searchable, organized
• File uploads — send PDFs, images, and documents for the AI to analyze
• Web search integration — optionally enable real-time web search in responses
• Multi-user support — create accounts for multiple people on your household or team
• Voice input and output — talk to your AI, hear responses read aloud
• Image generation — integrate with Stable Diffusion or DALL-E
• Custom system prompts — set permanent instructions for any model
• RAG (document chat) — upload files and ask questions about them

Prerequisites — What You Need Before Installing Open WebUI

• Ollama installed and running — if not, see: How to Install Ollama on Windows → | Mac → | Linux →
• At least one model downloaded — run ollama pull llama3.1 if you haven’t yet
• RAM: 4 GB free beyond what Ollama uses (Open WebUI itself is lightweight)
• Browser: Chrome, Firefox, Edge, or Safari — any modern browser works

Verify Ollama is running before proceeding:

ollama --version
ollama list

Method 1 — Install with Python pip (Easiest, No Docker)

This is the simplest method — no Docker required, works on Windows, Mac, and Linux.

Requirements: Python 3.11+

Check your Python version first:

python --version
# Needs to show Python 3.11 or higher

If you don’t have Python 3.11+, download it from python.org/downloads.

Install Open WebUI

pip install open-webui

Start Open WebUI

open-webui serve

Wait 30–60 seconds for it to initialize. Then open your browser and go to:

http://localhost:8080

You’ll see the Open WebUI setup screen. Create an admin account (this is local only — no external servers), and you’re in.

Windows tip: If open-webui isn’t recognized after install, use python -m open_webui serve instead, or close and reopen your terminal to reload PATH.

Method 2 — Install with Docker (Recommended for Servers)

Docker provides the most reliable and portable Open WebUI installation — ideal for home servers, Linux machines, and anyone who wants clean, isolated deployments.

If Ollama is installed directly on your machine

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

Open your browser → http://localhost:3000

If you want Ollama AND Open WebUI both in Docker

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

This single container runs both Ollama and Open WebUI together — useful for simple setups or NAS devices.

With NVIDIA GPU support in Docker

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

Method 3 — Install with Docker Compose

If you prefer Docker Compose (easier to manage and update), create a file called docker-compose.yml:

version: '3.8'
services:
  ollama:
    image: ollama/ollama
    container_name: ollama
    ports:
      - "11434:11434"
    volumes:
      - ollama_data:/root/.ollama
    restart: unless-stopped

  open-webui:
    image: ghcr.io/open-webui/open-webui:main
    container_name: open-webui
    depends_on:
      - ollama
    ports:
      - "3000:8080"
    environment:
      - OLLAMA_BASE_URL=http://ollama:11434
    volumes:
      - open_webui_data:/app/backend/data
    restart: unless-stopped

volumes:
  ollama_data:
  open_webui_data:

Then start everything with:

docker compose up -d

Access at http://localhost:3000. To update in the future: docker compose pull && docker compose up -d

First-Time Setup — Creating Your Admin Account

The first time you open Open WebUI in your browser, you’ll be prompted to create an account. This is a completely local account — no data goes anywhere external.

1. Enter your name, email address (any email works — it’s just a local identifier), and a password
2. Click Create Admin Account
3. You’re logged in — you’ll see the main chat interface

The first account created is automatically the admin. You can add more users later from Settings → Admin Panel → Users.

Connect Open WebUI to Ollama

Open WebUI automatically detects Ollama if both are running on the same machine. Verify the connection:

1. Click your profile icon (top right) → Admin Panel
2. Go to Settings → Connections
3. You should see http://localhost:11434 listed as the Ollama URL with a green checkmark

If the connection shows an error, make sure Ollama’s server is running:

ollama serve

How to Use Open WebUI — Complete Feature Guide

Switching Between AI Models

At the top of the chat window, click the model name dropdown. Every model you’ve downloaded with Ollama appears in this list. Switch instantly between Llama 3.1, Mistral, DeepSeek-R1, or any other installed model without closing the conversation.

You can also download new models directly from the Open WebUI interface:

1. Click your profile → Admin Panel → Settings → Models
2. Type a model name in the “Pull a model from Ollama.com” field
3. Click the download button — it pulls the model the same as ollama pull

Uploading Files and Documents

Click the paperclip icon (📎) in the chat input to upload files. Supported types:

• PDFs — upload and ask questions about the content
• Text files (.txt, .md, .csv) — analyze, summarize, or reformat
• Images — if you’re using a vision model like Llama 3.2 Vision or LLaVA
• Word documents (.docx) — extract and analyze content

Example: upload your 50-page PDF report and ask “Summarize the key findings in 5 bullet points.” The AI reads it entirely locally — no cloud processing, no data leaving your machine.

Web Search Integration

Enable real-time web search so the AI can look up current information:

1. Go to Admin Panel → Settings → Web Search
2. Select a search engine (DuckDuckGo works without an API key)
3. Enable the toggle
4. In chat, click the 🌐 globe icon to enable search for that conversation

Custom System Prompts

Set permanent instructions that apply to every conversation with a model:

1. Click your profile → Settings → General
2. Enter your system prompt in the “System Prompt” field
3. Example: “You are a helpful coding assistant. Always provide code examples. Be concise.”

You can also set model-specific system prompts by going to Workspace → Models and customizing each model’s settings independently.

Voice Input and Text-to-Speech

Click the microphone icon 🎤 in the chat input to speak your message. Open WebUI uses your browser’s speech recognition — works in Chrome and Edge without any extra configuration.

For AI voice responses: go to Settings → Interface → Text-to-Speech and enable it. The AI will read its responses aloud using your device’s built-in speech synthesis.

Creating and Sharing Conversations

• Export a chat: Click the three dots (…) next to any conversation → Export → Downloads as JSON or plain text
• Share a chat: Admin settings allow enabling shareable conversation links for multi-user setups
• Search conversations: Use the search bar in the sidebar to find any previous conversation by keyword

Open WebUI on a Network — Access from Any Device

One of the most powerful things about Open WebUI is running it on one computer and accessing it from any device on your home network — phone, tablet, other laptops, anything with a browser.

Step 1 — Find your server’s local IP address

# Windows
ipconfig | findstr "IPv4"

# Mac / Linux
ip addr show | grep "inet "
# Or: hostname -I

Note the IP — something like 192.168.1.42.

Step 2 — Allow Open WebUI through your firewall

# Windows (run as Administrator)
netsh advfirewall firewall add rule name="Open WebUI" protocol=TCP dir=in localport=8080 action=allow

# Linux with UFW
sudo ufw allow 8080/tcp

Step 3 — Access from other devices

On any device connected to the same WiFi network, open a browser and go to:

http://192.168.1.42:8080

(Replace with your actual server IP). You can now use your Ollama AI from your phone, tablet, or any other computer — all still processing locally on your main machine.

Keeping Open WebUI Updated

Update the pip version

pip install --upgrade open-webui

Update the Docker version

docker pull ghcr.io/open-webui/open-webui:main
docker stop open-webui
docker rm open-webui
# Then re-run the original docker run command

Update Docker Compose version

docker compose pull
docker compose up -d

Common Open WebUI Errors and Fixes

Error: “Could not connect to Ollama” in the interface

Ollama isn’t running. Open a terminal and start it:

ollama serve

Then refresh Open WebUI in your browser.

No models showing in the dropdown

You haven’t downloaded any Ollama models yet. In a terminal:

ollama pull llama3.1

Wait for the download to finish, then refresh Open WebUI — the model appears automatically.

Open WebUI is running but browser shows blank page

This is usually a port conflict or the server is still initializing. Wait 60 seconds and refresh. If still blank, check the terminal where you ran open-webui serve for error messages.

File uploads failing or timing out

Large file uploads can fail if the AI is already busy processing another request. Wait for the current generation to finish, then try uploading again. For very large PDFs (50+ pages), try splitting them into smaller chunks for better results.

Docker command not found on Windows

Install Docker Desktop for Windows from docker.com. Make sure WSL2 is enabled (Docker Desktop guides you through this). After installation, restart your terminal and Docker commands will work.

Frequently Asked Questions

Questions about your specific setup? Leave a comment below — I respond to every reader question about Open WebUI and Ollama.

About this guide: Tested with Open WebUI v0.5.x on Windows 11, macOS Sequoia, and Ubuntu 22.04. Both pip and Docker installation methods verified. Last updated March 2026.