TopPapersAI

Automatically get the best Computer Science papers from arXiv delivered to your email. Stay updated with top research without the noise.

Project Scope

TopPapersAI is a local-first research monitoring tool. It fetches recent arXiv papers, scores them with a configurable ranking strategy, stores subscribers locally, and sends digest emails. The project is intended as a practical automation tool rather than a hosted newsletter service.

Features

• Email notifications with top-rated papers
• 3 scoring methods: Rule-based, Transformer, or API (ChatGPT/Claude/Gemini)
• Easy setup - works in 5 minutes
• Privacy-focused - your data stays local

Quick Start

# Install
git clone https://github.com/Kevin-Li-2025/TopPapersAI.git
cd TopPapersAI
pip install -r requirements.txt

# Configure email (copy example and edit)
cp env_example.txt .env
# Edit .env with your email settings

# Register yourself
python3 cli.py register --email your@email.com --name "Your Name"

# Get papers
python3 cli.py notify

Main Commands

python3 cli.py register --email user@example.com    # Add new user
python3 cli.py notify                               # Send latest papers
python3 cli.py papers                               # View available papers
python3 cli.py users                                # List users
python3 cli.py scoring-methods                      # Compare scoring options

Architecture

cli.py              Command-line entry point
config.py           Runtime configuration and environment loading
paper_fetcher.py    arXiv query and paper normalization
scorer.py           Rule/API/model scoring strategies
email_sender.py     Digest rendering and SMTP delivery
templates/          Email templates
static/             Optional local web assets

Data and Privacy

• Subscriber data stays in the local project environment.
• Email credentials should be stored in .env, not committed.
• API-based scoring is optional; rule-based scoring can run without external model providers.

Email Setup (Gmail)

1. Enable 2-Factor Authentication
2. Generate App Password: Google Account Settings → Security → App Passwords
3. Update .env:

   EMAIL_FROM=your@gmail.com
   EMAIL_PASSWORD=your_16_char_app_password
   

Scoring Methods

Method	Speed	Cost	Accuracy
Rule-based	Fast	Free	Good
Transformer	Medium	Free	Better
API	Slow	Paid	Best

Switch anytime: python3 cli.py notify --method transformer

Configuration

Edit .env for your settings:

# Required
EMAIL_FROM=your@email.com
EMAIL_PASSWORD=your_app_password

# Optional  
SCORING_METHOD=rule-based    # rule-based, transformer, or api
MIN_SCORE_THRESHOLD=6.0      # Minimum score for "top" papers
UPDATE_FREQUENCY=24          # Hours between updates

arXiv Compliance

• Non-commercial use
• Proper attribution
• Respectful API usage

"Thank you to arXiv for use of its open access interoperability."

Troubleshooting

• Email not working? Check if you're using App Password (not regular password)
• No papers found? Try: python3 cli.py papers to see what's available
• Want better scores? Try: python3 cli.py notify --method transformer

---

Made for researchers. Star if this helps you stay updated with research.