Skip to main content

Your First App: Open WebUI

Let's deploy your first application with OnTree! We'll use Open WebUI - a powerful, self-hosted interface for running AI models locally. This tutorial will walk you through the entire process from start to finish.

What is Open WebUI?​

Open WebUI provides a ChatGPT-like interface for interacting with local LLMs (Large Language Models). It's perfect as a first app because it:

  • Has a beautiful, intuitive interface
  • Works immediately after deployment
  • Demonstrates OnTree's power in simplifying complex deployments

Step 1: Access the Templates Page​

  1. Open OnTree in your browser at http://localhost:8080
  2. Click "Create New App" on the dashboard
  3. You'll be taken to the templates page

Step 2: Select Open WebUI Template​

  1. Find "Open WebUI" in the template list
  2. Click "Use Template" to start the creation process

The template includes:

  • Pre-configured docker-compose.yml
  • Optimal resource settings
  • Persistent volume for your data
  • Automatic port allocation

Step 3: Configure Your App​

You'll see the app creation form with pre-filled values from the template.

App Name​

  • Default: openwebui
  • Your choice: Keep the default or choose something like ai-chat or my-assistant
  • Requirements: Lowercase letters, numbers, and hyphens only

Select an Emoji​

OnTree lets you personalize your apps with emojis:

  1. Browse the emoji picker showing 7 random emojis
  2. Click an emoji to select it (it will highlight in blue)
  3. Click the shuffle button 🔄 for different options
  4. Suggestion: 🤖 or 🧠 work great for AI apps!

Review Configuration​

The template automatically configures:

  • Image: ghcr.io/open-webui/open-webui:main
  • Port: Automatically allocated (usually 3000-4000 range)
  • Volume: Persistent storage at /app/backend/data

Step 4: Create and Start​

  1. Click "Create App" at the bottom of the form

  2. OnTree will:

    • Create the app directory
    • Generate the docker-compose.yml
    • Save your configuration

  3. On the app detail page, click the green "Create & Start" button

  4. Watch the operation logs as OnTree:

    • Pulls the Docker image
    • Creates the container
    • Starts Open WebUI

This may take 2-3 minutes on first run as the image downloads.

Step 5: Access Open WebUI​

Once the container is running:

  1. Look for the "Container Info" card
  2. Find the mapped port (e.g., 3000 → 8080)
  3. Click the port number or navigate to http://localhost:3000

First Time Setup​

When you first access Open WebUI:

  1. Create an account - This is your local admin account
  2. Configure your name and preferences
  3. You're ready to go!

Step 6: (Optional) Expose with a Domain​

If you've configured domains, you can make Open WebUI accessible at a custom URL:

  1. In the "Domain & Access" section, enter a subdomain:

    • Example: chat for chat.yourdomain.com
    • Or: ai for ai.yourdomain.com

  2. Click "Expose App"

  3. Access your app at https://chat.yourdomain.com with automatic HTTPS!

Using Open WebUI​

Adding AI Models​

Open WebUI supports multiple model backends:

  1. Ollama (Recommended for beginners):

    # Install Ollama on your host
    curl -fsSL https://ollama.com/install.sh | sh

    # Pull a model
    ollama pull llama2

  2. Configure in Open WebUI:

    • Go to Settings → Models
    • Add Ollama URL: http://host.docker.internal:11434
    • Your models will appear automatically

Features to Explore​

  • Chat Interface: Similar to ChatGPT
  • Multiple Models: Switch between different AI models
  • Conversation History: All chats are saved locally
  • Custom Prompts: Create reusable prompt templates
  • Document Upload: Chat with your documents (PDFs, text files)

Managing Your App​

Viewing Logs​

  1. Click "View Logs" on the app detail page
  2. See real-time container output
  3. Useful for troubleshooting

Stopping and Starting​

  • Stop: Click the "Stop" button when the container is running
  • Start: Click "Start" when the container is stopped
  • OnTree preserves your data between restarts

Editing Configuration​

  1. Click "Edit" in the Configuration card
  2. Modify the docker-compose.yml
  3. Click "Save" - OnTree will recreate the container if needed

Troubleshooting​

Container Won't Start​

Check the logs for errors:

  • Port conflict: The assigned port might be in use
  • Resource limits: Ensure you have enough RAM (2GB minimum)

Can't Access the Web Interface​

  1. Verify the container is running (green status)
  2. Check the mapped port in Container Info
  3. Try accessing via localhost instead of 127.0.0.1

Models Not Loading​

If using Ollama:

  1. Ensure Ollama is running on the host
  2. Use http://host.docker.internal:11434 as the Ollama URL
  3. Check that models are pulled: ollama list

What's Next?​

Congratulations! You've successfully deployed your first app with OnTree. Here's what to explore next:

Try More Templates​

  • Nginx: Simple web server for testing
  • PostgreSQL: Database server
  • Redis: In-memory data store

Learn OnTree Features​

Customize Open WebUI​

  • Add more AI models
  • Create custom prompts
  • Integrate with your workflow

Tips for Success​

  1. Start Simple: Get comfortable with basic operations before advanced features
  2. Check Logs: When something goes wrong, logs usually have the answer
  3. Use Emojis: They make finding apps in your dashboard much easier
  4. Regular Backups: OnTree preserves data, but backups are always wise

Welcome to the OnTree community! You're now ready to self-host any application with confidence. 🚀