OpenWebUI Deep Research Pipe

A powerful OpenWebUI pipe that integrates OpenAI's deep research models (o3-deep-research and o4-mini-deep-research) to provide comprehensive research capabilities with web search and code interpretation.

🚀 Features

• Deep Research Models: Access to OpenAI's specialized research models
• Web Search Integration: Automated web search for comprehensive research
• Code Interpreter: Built-in code execution for data analysis and visualization
• Background Processing: Long-running research tasks with progress tracking
• Prompt Enrichment: Automatic enhancement of research queries using GPT-4.1
• Flexible Configuration: Extensive customization options via Valves
• Fallback Support: Graceful fallback to standard models when needed
• Test Mode: Built-in diagnostics for troubleshooting

⚠️ Important Requirements

OpenAI API Access: The deep research models (o3-deep-research and o4-mini-deep-research) require special access from OpenAI. Regular API keys may not have access to these models. If you're receiving empty responses, your API key likely doesn't have the required permissions.

📋 Installation

1. Download the OpenWebUi Function v1.py file
2. In OpenWebUI, go to Admin Panel → Functions
3. Click "+" to add a new function
4. Upload or paste the code from OpenWebUi Function v1.py
5. Configure the Valves (see Configuration section below)

⚙️ Configuration

The pipe uses a comprehensive Valves system for configuration. Access these settings in OpenWebUI under the function's settings:

Essential Settings

Setting	Default	Description
API_KEYS	""	Required: Your OpenAI API key(s), comma-separated for multiple keys
BASE_URL	"https://api.openai.com/v1"	OpenAI API base URL
NAME_PREFIX	"Deep Research: "	Prefix for model names in the UI

Research Features

Setting	Default	Description
ENABLE_WEB_SEARCH	True	Enable web search capabilities
ENABLE_CODE_INTERPRETER	True	Enable code execution for analysis
ENABLE_PROMPT_ENRICHMENT	False	Use GPT-4.1 to enhance research queries
ENABLE_CLARIFICATION	False	Ask clarifying questions before research
MAX_TOOL_CALLS	None	Limit tool usage to control costs

Performance & Reliability

Setting	Default	Description
USE_BACKGROUND_MODE	True	Recommended: Use background processing for reliability
POLL_INTERVAL	3	Seconds between status checks
MAX_POLL_ATTEMPTS	600	Maximum polling attempts (30 minutes)
REQUEST_TIMEOUT	120	API request timeout in seconds
CONNECTION_TIMEOUT	30	Connection timeout in seconds

Troubleshooting

Setting	Default	Description
TEST_MODE	False	Enable diagnostic mode
USE_FALLBACK_MODEL	False	Fall back to standard models if deep research fails
FALLBACK_MODEL	"gpt-4-turbo-preview"	Model to use as fallback

🎯 Usage

Basic Research Query

What are the latest developments in quantum computing in 2024?

Complex Analysis Request

Analyze the economic impact of renewable energy adoption in the EU over the past 5 years, including market trends, policy effects, and future projections. Please include relevant data visualizations.

Technical Research

Compare the performance characteristics of different machine learning frameworks for large language model training, focusing on memory efficiency and scalability.

🔧 Available Models

The pipe provides access to two specialized research models:

• Deep Research: o3-deep-research - Advanced research capabilities
• Deep Research: o4-mini-deep-research - Lightweight research model

🛠️ Troubleshooting

Common Issues

Empty or No Response

• Cause: API key lacks access to deep research models
• Solution: Contact OpenAI to request access, or enable USE_FALLBACK_MODEL

Timeout Errors

• Cause: Deep research can take several minutes
• Solution:
  • Ensure USE_BACKGROUND_MODE is enabled
  • Increase REQUEST_TIMEOUT and CONNECTION_TIMEOUT
  • Try during off-peak hours

Authentication Errors

• Cause: Invalid or missing API key
• Solution:
  • Verify API key is correct in API_KEYS
  • Enable TEST_MODE to verify API key validity

Diagnostic Mode

Enable TEST_MODE in Valves to:

• Test API key validity
• Verify model access
• Debug connection issues
• View detailed request/response information

📊 How It Works

Background Processing Flow

1. Query Preparation: Optionally enriches user input with GPT-4.1
2. Background Creation: Submits research task to OpenAI's background API
3. Status Monitoring: Polls for completion with regular status updates
4. Result Streaming: Streams final results as they become available

Tool Integration

The pipe automatically configures research tools based on your settings:

• Web Search: Accesses current information from the internet
• Code Interpreter: Executes Python code for data analysis and visualization
• Reasoning Summary: Provides structured thinking process

🔒 Security & Privacy

• API keys are handled securely and never logged
• Research queries are processed through OpenAI's systems
• Background mode stores research temporarily on OpenAI's servers
• Enable store: true for optimal performance

📈 Performance Tips

1. Use Background Mode: Always keep USE_BACKGROUND_MODE enabled for reliability
2. Multiple API Keys: Add multiple keys separated by commas for better rate limits
3. Timeout Configuration: Adjust timeouts based on your research complexity
4. Tool Limits: Use MAX_TOOL_CALLS to control costs for expensive research

🤝 Contributing

Contributions are welcome! Please feel free to:

• Report issues
• Suggest improvements
• Submit pull requests
• Share usage examples

📄 License

MIT License - see the code header for full details.

🙏 Acknowledgments

• OpenAI for the deep research models and API
• OpenWebUI team for the extensible pipe system
• Contributors and users providing feedback

📚 Additional Resources

• OpenAI API Documentation
• OpenWebUI Documentation
• Deep Research Models Guide

---

Note: This pipe requires OpenAI API access and special permissions for deep research models. Regular API keys may not have access to o3-deep-research and o4-mini-deep-research models.