Zotero-arXiv- Daily - Personal Assistant for Scientific Research
Sound familiar: you want to stay on top of the latest scientific publications, but scrolling through hundreds of arXiv papers every day is a luxury you can't afford? That's where Zotero-arXiv- Daily comes in — a project that automates finding relevant research based on your preferences.
How does it work?
The project analyzes your Zotero library (a popular reference manager) and sends you a daily email with fresh arXiv papers that might interest you. And it does all of this using GitHub Actions — completely free.
Key features:
- Personalized recommendations — the system considers your current research interests based on papers in your Zotero library
- AI assistant — each paper gets a brief summary (TL;DR) generated by an LLM
- Fully automated — set it up once, and receive daily emails
- Free to use — runs within GitHub Actions' free tier
- Flexible settings — you can specify particular paper categories and adjust how many recommendations you receive
Getting started
Setup takes just 5 minutes:
- Fork the repository (and give the author a star 😊)
- Configure the required variables in GitHub Actions settings
- Enjoy your daily paper selections
What needs to be configured?
Key settings include:
- Your Zotero account ID and API key
- arXiv paper categories (e.g., cs.AI for artificial intelligence)
- SMTP server details for sending emails
- Your email address for notifications
Technical details
Under the hood, the project uses:
- Zotero API to analyze your library
- arXiv API to fetch new publications
- Sentence Transformers to compute semantic similarity between papers
- LLM (Qwen2.5-3b-instruct) to generate summaries
All processing happens directly in GitHub Actions, so nothing needs to be installed locally.
Who is this for?
- PhD students and researchers
- Engineers tracking cutting-edge research
- Instructors wanting to incorporate current materials into their courses
- Anyone tired of manually monitoring new publications
Limitations
- The recommendation algorithm is fairly basic — it only considers semantic similarity of abstracts
- Generating a TL;DR for a single paper takes roughly 70 seconds
- The free GitHub Actions tier caps at 2000 minutes monthly
Zotero-arXiv- Daily is ideal for researchers looking to streamline their literature discovery process. The project continues gaining traction with over 3k GitHub stars and regular updates. If you regularly work with academic papers, this tool is worth trying!
Progetti correlati