Inbox Organizer Agent MVP

An AI-powered email management tool that automatically categorizes emails, flags important messages, and saves summaries to Notion.

Features

• 📧 Email Reading: Connects to Gmail via API
• 🏷️ Auto Categorization: Sorts emails into Work, Personal, Promotions, Updates
• ⚡ Important Email Detection: Flags urgent/important emails
• 📝 AI Summarization: Generates concise summaries using OpenAI
• 📊 Notion Integration: Saves important emails to Notion database
• 🖥️ Web Interface: User-friendly Streamlit dashboard

Quick Setup

1. Install dependencies:

   pip install -r requirements.txt

2. Configure environment:

   cp .env.example .env
   # Edit .env with your API keys

3. Gmail Setup:

   • Go to Google Cloud Console
   • Enable Gmail API
   • Download credentials.json
   • Place in project root

4. Notion Setup:

   • Create integration at notion.so/my-integrations
   • Create database with properties: Subject (Title), Sender (Text), Category (Select), Date (Text), Summary (Text), Important (Checkbox)
   • Get database ID from URL

5. Run the app:

   streamlit run app.py

Usage

Manual Processing (Web Interface)

1. Open the web interface: streamlit run app.py
2. Click "Process Emails" to analyze recent emails
3. View categorization results and important emails saved to Notion

Automated Processing (Background)

Option 1: Simple Scheduler

python run_scheduler.py  # Runs 3x daily (9AM, 2PM, 6PM)

Option 2: Daemon Service (Recommended for hosting)

python daemon.py start   # Start background service
python daemon.py stop    # Stop background service

• Processes emails 3 times daily automatically
• Runs independently even when app is closed
• Check logs in email_processing_log.json
• Perfect for server deployment

Configuration

Edit config.py to customize:

• Email categories
• Important keywords
• Processing parameters

MVP Limitations

• Processes recent emails only (not historical)
• Basic categorization (4 categories)
• Simple importance detection
• No email actions (mark as read, archive, etc.)

Next Steps

• Add more sophisticated categorization
• Implement email actions
• Add scheduling/automation
• Improve error handling
• Add user feedback loop