The traditional literature review strategy — search keywords, find papers, check their references, chase references of references — misses the network structure of academic knowledge. ResearchRabbit and Semantic Scholar make that network visible and navigable.
Semantic Scholar
Semantic Scholar (semanticscholar.org) is a free academic search engine built by the Allen Institute for AI. Its distinctive features: citation counts that show which papers in a field are most influential, “highly influential citations” (papers where a given paper was central to the citing paper’s argument), and an AI-powered “TLDR” one-sentence summary of each abstract.
The Semantic Reader feature lets you read PDFs with inline citation context — hover over a citation to see a snippet of what the referenced paper actually says, without opening it. For papers with 50+ citations, this speeds up reading considerably.
ResearchRabbit
ResearchRabbit (researchrabbit.ai) builds visual citation networks. Start with one or two seed papers central to your research question, and it generates a network diagram showing: papers that cite your seeds (later work building on them), papers your seeds cite (foundations), and related papers with similar citation patterns (peers). The visualization reveals clusters, bridges between subfields, and seminal papers that keep appearing across multiple nodes.
A Practical Discovery Workflow
Step 1: Find 2–3 central papers in your topic through Google Scholar or Semantic Scholar. Step 2: Add them to ResearchRabbit and explore the citation network. Step 3: Identify the 5–10 most connected papers — these are usually the papers every researcher in the field has read. Step 4: Import those core papers into Semantic Scholar to check their “highly influential” citation chains. Step 5: Export the full list to Zotero and tag by relevance.
This workflow regularly surfaces papers that direct keyword search misses, because papers within the same citation cluster often use different terminology.
Connected Papers
Connected Papers (connectedpapers.com) offers a similar visualization to ResearchRabbit but with a different algorithm — it’s based on co-citation and bibliographic coupling rather than direct citations. The two visualizations complement each other: ResearchRabbit shows citation lineage; Connected Papers shows papers that are “read together” by researchers even without direct citation links.
