Gmail LLM

Using GPT4 to extract interesting links from my daily newsletters

Why? I spend too much time opening and checking through all my newsletters subscriptions, would rather have all interesting links in one page.

Inspired by isafulf/inbox_cleaner

Images

Prerequisites

• Python 3.10 or higher
• Gmail account
• Google Cloud account with Gmail API enabled
• OpenAI API key

Setup

1. Clone repo:

   git clone https://github.com/benthecoder/gmail_llm.git
   cd gmail_llm
   

2. Install packages:

   pip install -r requirements.txt
   

3. Set up Google API credentials:

   • create a new OAuth 2.0 Client ID (instructions)
   • Download JSON file and rename it to credentials.json.
   • Put credentials.json in the gmail_llm directory.

4. Set OpenAI API key:

   • Get an OpenAI API key here
   • Rename .env.local to .env and set your key

   OPENAI_API_KEY=<YOUR_API_KEY>
   

Usage

Run the script:

python main.py

see an example in summary folder

TODO

• improve prompt to extract as many links as possible and have better categorization
• Experiment with different forms of output, maybe article form? Have it copy someone's writing style, publish it to substack, could be a newsletter automation with AI thing
• Check if text extracted from emails can be cleaned up, any redundant info fed into prompt?
• Run this on cloud/local and dump results in DB so I can read into my blog or in notion

Links

• Docs for listing emails users.messages.list
• Docs for querying Searching for Messages
• Counting tokens with tiktoken