Smart Facts Collector

A mobile-optimized web application that uses GPT-4o to intelligently extract facts from unstructured text and submit them to the OMI API.

Features

• GPT-4o AI-Powered Extraction: Uses advanced AI to identify and extract meaningful facts from messy, unstructured text
• Mobile-optimized UI with AI toggle
• Multi-line text input supporting various formats
• Smart fact extraction from complex text
• Submission of each fact to the OMI API
• Detailed status updates and error handling

Setup

1. Install the required dependencies:

pip install -r requirements.txt

2. Set your OpenAI API key (required for AI-powered extraction):

# On Linux/macOS
export OPENAI_API_KEY=your_openai_api_key_here

# On Windows
set OPENAI_API_KEY=your_openai_api_key_here

Alternatively, you can edit the app.py file and replace the placeholder with your API key:

OPENAI_API_KEY = os.environ.get("OPENAI_API_KEY", "your_openai_api_key_here")

3. go to Omi AI app in Appstore, create a new app with "external integration capability", and allow "facts"

Go to settings, copy your App ID and create private key

4. in app.py file, provide the app ID and private key you've generated

5. Run the Flask application:

python app.py

6. Open your browser and navigate to http://localhost:5001

How to Use

1. Toggle AI-powered extraction on or off (on by default)
2. Enter text in the textarea - this can be:
   • Notes from a book or meeting
   • Personal learning journal entries
   • Bullet points of insights
   • Any unstructured text containing facts
3. Click the "Extract & Submit Facts" button
4. The application will:
   • Use GPT-4o to intelligently identify facts (if enabled)
   • Fall back to rule-based extraction if AI is disabled or unavailable
   • Display the extracted facts
   • Submit each fact to the OMI API
   • Show the results of the submission

Technical Details

• The frontend is built with HTML, CSS, and JavaScript
• The backend uses Flask to serve the application and handle API requests
• GPT-4o is used for intelligent fact extraction
• Rule-based extraction serves as a fallback
• Facts are sent to the OMI API individually
• Basic rate limiting is implemented to prevent API overload

Troubleshooting

If you encounter port conflicts:

1. You can change the port in app.py by modifying the line: app.run(host='0.0.0.0', port=5001, debug=True)
2. On macOS, port 5000 is often used by AirPlay Receiver. You can disable this service in System Preferences > Sharing.

If the AI extraction is not working:

1. Make sure you've set the OpenAI API key correctly
2. Check the terminal for any API errors
3. Try toggling to rule-based extraction as a fallback