LitBeacon

How to use LitBeacon

Last updated 24 October 2026

LitBeacon helps you find your way around an unfamiliar research field. Type the name of a field and LitBeacon gathers about two dozen real papers that you can look at three ways: a guide that suggests where to start reading, a timeline of how the topic developed, and a map of how the papers connect.

Search for a specific field or topic, such as graph neural networks, CRISPR gene editing or Arabidopsis TRAPPII. LitBeacon then:

  1. collects about 150 candidate papers from OpenAlex in several ways (by relevance, citations, recency, reviews, and what the top papers cite);
  2. asks an AI model to read each candidate's title and abstract and say whether it is about your topic, near it, or about something else, quoting the passage it relied on;
  3. chooses about 24 papers that together span the field, and places each in one reading category;
  4. connects them by citations, shared references, similar content, shared topics and shared authors.

A fresh search takes a little while, since the AI model needs to process each candidate paper. If you want to see a quick example, click on one of the Examples on the start page.

The reading categories

In the Diagram, read from left to right; in the Cards list, from top to bottom. Hover over, or tab to, a category name in the guide to see its explanation.

Subtopic names come from the papers' OpenAlex topics and titles. A subtopic with only one paper in a section is listed under Other topics. Click on the paper to see the exact subtopic name.

You can also change any paper's category: choose it when you add the paper, or use Move category to… in the paper's window.

Working with papers

Three layouts: Guide, Timeline and Map

A node is a circle or diamond in the map and represents a paper. A circle usually marks an experimental paper, while a diamond marks a review or survey. The node size follows the citation measure chosen in Node size: total citations, or citations per year. Scroll inside the map to zoom in or out. Click on the arrows on the edge of the left and right sidebars to expand or collapse them.

What the arrows between nodes mean

An arrow between two nodes indicates that the two papers are related. By default, the arrow points from the cited paper to the citing paper (from old to new). The direction can be reversed with the Arrow direction toggle on the right sidebar. The style of the arrow indicates the type of relationship: one paper cites the other, they share references, their content is similar, they share topics, or they share an author. Only the strongest few connections of each paper are drawn.

In the bar above the map, Show links hides the weaker connections, and Connections chooses which kinds are shown. Unticking a category will hide its connections, except citations, which stay as gray arrows.

What LitBeacon does not claim

The guide generated by LitBeacon is just a starting point, not a complete or expert-checked reading list. Use it to get oriented, then follow the references of the papers you find useful. It's always a good idea to double-check the information by asking someone who knows the field.

Which fields it suits

LitBeacon works best in fields whose literature is mostly journal articles, reviews and preprints with abstracts, as in most of the life sciences, natural sciences, medicine and computer science. Its testing so far has been in those areas, mostly biology.

Questions

Why only about 24 papers?

A reading guide is useful because it is short. You can add papers to a map, up to 40, and follow any paper's references from its link.

Why are there fewer than three Starters?

Starters must be reviews or tutorials about your topic. When the candidates include fewer than three, LitBeacon shows fewer rather than filling the slots with research papers. You can add one yourself with the + button next to the Starters category title.

Why did a similar search give different papers?

The databases match words, so "genome editing" and "gene editing" find partly different candidates. Try both and compare.

What is sent to the AI model?

Your search term and each candidate paper's title and abstract, never your account. The privacy note has the details.

Why did my search come back empty?

Either no relevant papers were found, or the AI relevance check could not run. Since LitBeacon uses an AI model to assess relevance, it has a daily limit of new searches. Please write to litbeacon.feedback@gmail.com if you ran out of limits but need to do a new search.

How many guides can I make?

A guest can build 2 guides and add papers to a guide 4 times. A verified account can build 10 guides and add papers 20 times each day, counted again from midnight UTC. Until its email is confirmed, an account has a guest's allowance.

Something in my guide is wrong. Where do I say so?

Please write to litbeacon.feedback@gmail.com with the search and the paper. Reports like this are how the guide gets better.