Open WebUI

Feature-rich self-hosted AI interface integration guide

Open WebUI is a feature-rich self-hosted AI platform that supports fully offline operation. Through APIYI, you can integrate various mainstream large language models in Open WebUI.

Quick Deployment

Docker Quick Start

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

Docker Compose Deployment

version: '3.6'

services:
  open-webui:
    image: ghcr.io/open-webui/open-webui:main
    container_name: open-webui
    ports:
      - "3000:8080"
    volumes:
      - open-webui:/app/backend/data
    environment:
      - OPENAI_API_BASE_URL=https://api.apiyi.com
      - OPENAI_API_KEY=Your APIYI key
    restart: unless-stopped

volumes:
  open-webui:

Configure APIYI

Method 1: Environment Variable Configuration

Set environment variables during deployment:

docker run -d -p 3000:8080 \
  -e OPENAI_API_BASE_URL=https://api.apiyi.com \
  -e OPENAI_API_KEY=Your APIYI key \
  -v open-webui:/app/backend/data \
  --name open-webui \
  ghcr.io/open-webui/open-webui:main

Method 2: Interface Configuration

1. Access Open WebUI admin interface
2. Go to Settings > Connections
3. Configure in OpenAI API section:
   • API Base URL: https://api.apiyi.com/v1
   • API Key: Enter your APIYI key
4. Click save configuration

Configuration Key Points

• API Base URL needs to include the /v1 suffix
• API Key can be obtained from APIYI Console
• Recommended to use environment variable method for easier management and updates

Supported Models

Open WebUI supports the following model series through APIYI:

Recommended Models

Model Series	Model ID	Features
GPT-4 Turbo	gpt-4-turbo-2024-04-09	Latest GPT-4, balanced performance
Claude Sonnet	claude-sonnet-4-20250514	Long text processing, creative writing
Gemini Pro	gemini-2.5-pro	Multimodal capabilities, fast response
GPT-3.5 Turbo	gpt-3.5-turbo	Economical and practical, daily conversation

Featured Function Models

Feature	Recommended Model	Description
Vision Understanding	gpt-4-vision-preview	Image analysis and understanding
Code Generation	gpt-4-turbo	Programming assistance and code optimization
Reasoning Thinking	claude-sonnet-4-20250514-thinking	Shows thinking process

Core Features

RAG (Retrieval-Augmented Generation)

Open WebUI supports document upload and knowledge base features:

1. Document Upload

   • Supports PDF, TXT, DOCX and other formats
   • Automatic vectorization storage
   • Supports multilingual documents

2. Knowledge Base Management

   • Create specialized knowledge bases
   • Document classification and tagging
   • Intelligent retrieval matching

OpenAI Compatible API

Open WebUI provides a complete OpenAI compatible API:

# Chat completion
curl -X POST "http://localhost:3000/api/chat/completions" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer Your APIYI key" \
  -d '{
    "model": "gpt-4-turbo",
    "messages": [
      {"role": "user", "content": "Hello, world!"}
    ]
  }'

Tool Integration

Supports external tools and plugins:

• Web search
• Code execution
• Image generation
• Document processing

Advanced Configuration

Multi-Model Configuration

Configure multiple model sources in docker-compose.yml:

environment:
  - OPENAI_API_BASE_URL=https://api.apiyi.com
  - OPENAI_API_KEY=Your APIYI key
  - ENABLE_OPENAI_API=true
  - ENABLE_OLLAMA_API=false

User Permission Management

environment:
  - ENABLE_SIGNUP=false
  - DEFAULT_USER_ROLE=user
  - WEBHOOK_URL=Your webhook address

Data Persistence

volumes:
  - open-webui:/app/backend/data
  - ./uploads:/app/backend/data/uploads
  - ./vector_db:/app/backend/data/vector_db

API Integration Examples

Python Integration

import requests

# Open WebUI API endpoint
api_url = "http://localhost:3000/api/chat/completions"

# Request configuration
headers = {
    "Content-Type": "application/json",
    "Authorization": "Bearer Your APIYI key"
}

data = {
    "model": "gpt-4-turbo",
    "messages": [
        {"role": "user", "content": "Explain the basic principles of quantum computing"}
    ],
    "stream": False
}

# Send request
response = requests.post(api_url, headers=headers, json=data)
result = response.json()
print(result["choices"][0]["message"]["content"])

JavaScript Integration

const apiUrl = 'http://localhost:3000/api/chat/completions';

const requestData = {
  model: 'gpt-4-turbo',
  messages: [
    { role: 'user', content: 'Write a simple Python function' }
  ]
};

fetch(apiUrl, {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json',
    'Authorization': 'Bearer Your APIYI key'
  },
  body: JSON.stringify(requestData)
})
.then(response => response.json())
.then(data => {
  console.log(data.choices[0].message.content);
});

Troubleshooting

Common Issues

Connection Failed

• Check if API Base URL is correct: https://api.apiyi.com/v1
• Verify API Key validity
• Confirm firewall settings

Model Unavailable

• Check account balance
• Confirm model is within service scope
• Check APIYI service status

Upload Failed

• Check file format support
• Confirm sufficient storage space
• Verify file size limits

Log Debugging

Enable debug mode:

docker logs -f open-webui

View detailed logs:

environment:
  - LOG_LEVEL=DEBUG
  - WEBUI_DEBUG=true

Best Practices

Performance Optimization

1. Model Selection

   • Daily conversation use GPT-3.5 Turbo
   • Complex tasks use GPT-4 Turbo
   • Long text processing use Claude Sonnet

2. Caching Strategy

   • Enable conversation caching
   • Set reasonable cache expiration time
   • Regularly clean unused cache

3. Resource Management

   • Monitor memory usage
   • Set reasonable concurrent limits
   • Regularly backup user data

Security Configuration

environment:
  - ENABLE_ADMIN_EXPORT=false
  - ENABLE_ADMIN_CHAT_ACCESS=false
  - JWT_EXPIRES_IN=7d

Monitoring Alerts

Integrate monitoring system:

environment:
  - ENABLE_WEBHOOKS=true
  - WEBHOOK_URL=https://your-monitoring-url

Need more help? Please check Open WebUI Official Documentation or visit APIYI Official Website.