Open WebUI + Ollama: Complete Setup Guide (2026)

作者： Jakub Rusinowski · 最后更新： 2026年7月21日

Open WebUI is the most popular graphical interface for Ollama. It gives you a ChatGPT-style web interface running entirely on your own hardware, with no data leaving your machine.

本指南内容

• What You'll Build
• Prerequisites
• Step 1: Install Docker
• Step 2: Start Ollama
• Step 3: Launch Open WebUI with Docker
• Step 4: First Launch
• Step 5: Access From Other Devices
• Configure RAG (Chat with Your Documents)

Open WebUI is the most popular graphical interface for Ollama. It gives you a ChatGPT-style web interface running entirely on your own hardware, with no data leaving your machine.

What You'll Build

A fully offline, private AI chat interface accessible from any device on your local network. Features include:

• Chat history stored locally
• Multiple models switchable mid-conversation
• Image uploads (with vision models)
• RAG (upload documents to chat with)
• User management (useful for home/office shared access)
• OpenAI-compatible API proxy

Prerequisites

• Ollama installed (see the Ollama setup guide)
• Docker Desktop installed (Windows/macOS) or Docker Engine (Linux)
• 8 GB+ RAM, any GPU (or CPU-only)

Step 1: Install Docker

Windows/macOS: Download Docker Desktop from docker.com and follow the installer.

Linux (Ubuntu):

curl -fsSL https://get.docker.com | sh
sudo usermod -aG docker $USER

Step 2: Start Ollama

Make sure Ollama is running and accessible:

ollama serve

Pull your first model:

ollama pull llama3.1:8b

Step 3: Launch Open WebUI with Docker

If Ollama is running on the same 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

If Ollama is on a different machine (e.g., desktop accessed from laptop):

docker run -d -p 3000:8080 \
  -e OLLAMA_BASE_URL=http://192.168.1.100:11434 \
  -v open-webui:/app/backend/data \
  --name open-webui \
  --restart always \
  ghcr.io/open-webui/open-webui:main

Step 4: First Launch

1. Open your browser to http://localhost:3000 2. Create an admin account (the first user becomes admin) 3. Select a model from the dropdown at the top 4. Start chatting!

Step 5: Access From Other Devices

To use your AI from your phone or tablet on the same network: 1. Find your computer's local IP: run ipconfig (Windows) or ip addr (Linux/Mac) 2. Open http://192.168.x.x:3000 on your phone 3. Works just like a web app — no app install required

Configure RAG (Chat with Your Documents)

Open WebUI has built-in RAG support: 1. Go to Settings → Documents and set the embedding model to nomic-embed-text 2. Pull the embedding model: ollama pull nomic-embed-text 3. In any chat, click the + button and upload a PDF, Word doc, or text file 4. Ask questions about the document — Open WebUI retrieves relevant chunks and feeds them to the LLM

Useful Settings to Configure

Setting	Location	Recommendation
Default model	Settings → Interface	Set your most-used model
Title generation	Settings → Interface	Auto-generate conversation titles
RAG embedding	Settings → Documents	Use nomic-embed-text
Memory	Settings → Personalization	Enable for persistent user context
System prompt	Model → System Prompt	Customize the AI's personality

Updating Open WebUI

docker pull ghcr.io/open-webui/open-webui:main
docker stop open-webui
docker rm open-webui
# Re-run the docker run command from Step 3

Your chat history is stored in the Docker volume (open-webui) and persists across updates.

Troubleshooting

"Connection error" when opening WebUI:

• Make sure Ollama is running: ollama serve
• Check the --add-host flag is included in your Docker command

Models not showing up:

• Run ollama list to confirm models are pulled
• Refresh the WebUI model selector

Slow responses:

• Ensure your GPU is being used: look for GPU activity in Task Manager / nvtop
• Check that Ollama is not running in CPU-only mode

→ Back to all guides | → Check hardware compatibility

← 所有指南 | 检查 GPU 兼容性