Citation Graph gives you a visual way to explore how research papers connect to each other.
Instead of running a keyword search, you start with papers you already trust (called seeds) and discover related work based on citation relationships.
The papers surfaced in the graph are meaningfully connected to your seeds, either directly cited by them, cited alongside them, or closely related through shared references.
This helps you:
Discover relevant papers you may have missed
Identify foundational and follow-up research
Understand how research evolved over time
Spot clusters and potential gaps in the literature
How to use Citation Graph
Step 1: Open Citation Graph
You can access it in two ways:
Click Graph in the left sidebar
Click the Citation Graph button on any paper’s sidebar (this automatically seeds the graph with that paper)
Step 2: Add Seed Papers
Citation Graph requires at least one seed paper to generate results.
You can add up to 5 seed papers
Seeds can be added from:
Recently saved papers (shown by default on the empty Graph page)
The search bar on the Graph page (powered by Quick Search)
Simply search for a paper and add it as a seed.
Step 3: Choose how many papers to show
Use the Density button at the bottom of the graph to choose how many papers appear.
You can view:
20 papers
40 papers
60 papers
Choose a lower density for a simpler, more focused graph. Choose a higher density to see more connected papers and explore a broader research area.
Step 4: Explore the Graph
Once generated, the graph displays papers positioned based on their citation relationships to your selected seeds.
You can:
Promote a result to a seed to refine and regenerate the graph
Exclude papers to remove them from the network
Hover over papers to see citation connections
Click into papers to review more details
Continue iterating to explore deeper connections
Seed papers and their 1-hop neighbors are color-coded, making it easier to see how each seed’s citation network branches through the graph.
Step 5: Chat with the Graph
At the bottom of the Graph page, you’ll see a Chat button. All seed papers and graph results are automatically attached as context.
Ask how specific papers are connected
Explore how research built on earlier work
Identify patterns or trends
Note: Citation Graphs are not saved in beta. Closing or refreshing the page will clear the graph.
How citation graph ranking works
The graph is built from the papers you select as seeds. All results are ranked based on their relevance to those seed papers.
Relevance is determined using three citation-based signals:
Direct citation relationships
The paper directly cites your seed, or your seed cites it.
Co-citation relationships
How often the paper and your seed appear together in other papers’ references.
Bibliographic coupling
How many references a paper shares with your seed papers (i.e., whether they cite similar prior work).
Together, these signals surface papers that are closely connected to your selected seeds through citation relationships.
Unlike a traditional citation crawl (which looks only at references forward or backward from one paper), the Citation Graph combines multiple citation signals to help you discover related work you might otherwise miss.
How to interpret the graph
The graph uses position, size, color, opacity, and arrows to help you understand relationships at a glance.
You can also refer to the legend in the top-right corner of the graph for a quick explanation of node colors, size, opacity, and ranking signals.
Horizontal axis = time
Papers on the left are older. Papers on the right are more recent.
Vertical axis = citation strength
Papers positioned higher in the map generally have higher citation counts.
Note: Vertical spacing is approximate and may not perfectly reflect citation count due to layout and space constraints.
Node size = citation count
The size of each paper node represents its citation count.
Larger nodes mean more citations.
Color = seed clusters
Each seed paper has a color. Papers that are 1 hop away from a seed appear in that seed’s color.
Gray nodes are bridge papers that help connect areas of the graph but are not directly connected to a seed.
Opacity = connectedness
Well-connected papers appear at full opacity.
Loosely connected papers appear more faded.
This helps you quickly spot papers that play an important role in the graph
Arrows = citation direction
Arrows represent citation relationships.
An arrow points from a paper to the paper that it cites.
Default arrows show connections between seed papers.
When you hover over a paper, the graph highlights citation connections associated with that paper.
This means you can visually trace:
Which papers influenced others
Which papers built on earlier work
How research flows over time
Where clusters or bridges appear in the literature
FAQs
Can I save my citation graph?
Not yet. Graphs are not saved in beta. Closing or refreshing the page will clear the graph.
How many seed papers can I use?
You can add up to 5 seed papers.
How many papers can I view in a graph?
You can choose to view 20, 40, or 60 papers in a graph.
Does Citation Graph use keyword matching?
No. It ranks papers based on citation relationships to your selected seeds.
What do the colors mean in my citation graph?
Seed papers and their 1-hop neighbors are color-coded so you can see how each seed’s citation network branches through the graph.
Can I refine my graph after generating it?
Yes. You can:
Promote results to new seeds (+ Seed button)
Exclude papers (minus button)
Regenerate the graph to explore further