Newsletter Digest Agent

An AI-powered agent that monitors your Gmail for newsletters, extracts key insights using Google Gemini, and compiles them into a Notion database and Daily Summary page.

NOTE: This agent was built via AI. Design brainstorm and PRD via Claude Sonnet. Code, testing, and utilities via Gemini/Antigravity.

Features

• Automated Search: Scans Gmail for newsletters based on your filters.
• AI Extraction: Uses Gemini Flash to extract summaries, key takeaways, and categories from emails.
• Notion Integration: Creates individual article pages and a consolidated "Daily Summary" in Notion.
• Chunking Support: Handles large summaries by splitting content into Notion-compatible blocks.
• Deduplication: Tags emails as processed to prevent duplicate entries.
• Agent Icon: assets/Newsletter_agent.jpg can be used when creating Notion API Connection (and will be used as the author avatar in Notion). The image was generated by Antigravity/Nano Bananna.

Setup

1. Prerequisites:

   • Python 3.10+
   • Google Cloud Project (Gmail API enabled)
   • Google Gemini API Key
   • Notion Integration Token

2. Environment Variables: Create a .env file:

   GEMINI_API_KEY=your_key_here
   NOTION_API_KEY=your_key_here
   NOTION_ARTICLES_DB_ID=your_db_id
   NOTION_SUMMARY_DB_ID=your_db_id

3. Authentication:

   • Place credentials.json (Gmail OAuth) in the root.
   • Run python gmail_auth.py to generate token.json.

4. Install Dependencies:

   pip install -r requirements.txt

Usage

Run the agent manually:

python main.py

Scheduling

Uses launchd on macOS for daily execution at 8:00 AM.

1. Generate the configuration file from the template:

   sed "s|{{PROJECT_DIR}}|$PWD|g" com.ben.newsletter_agent.plist.template > com.ben.newsletter_agent.plist

2. Install the agent:

   cp com.ben.newsletter_agent.plist ~/Library/LaunchAgents/
   launchctl load ~/Library/LaunchAgents/com.ben.newsletter_agent.plist

Project Structure

• main.py: Entry point and orchestration.
• gmail_client.py: Handles Gmail API searching and label management.
• llm_processor.py: Interacts with Gemini API for content extraction.
• notion_agent.py: Manages Notion pages and databases.