Connected Papers.

Connected Papers is a web-based discovery tool for getting oriented in an academic field. Give it a paper, DOI, arXiv, Semantic Scholar or PubMed URL, or a title, and it builds a visual graph of related papers. That makes a new topic easier to map, but it does not replace a systematic search strategy, full-text screening, or checking the original sources.

Who it is for

The tool suits students, researchers, librarians, and R&D teams that want to move from a known paper to relevant neighbors. It is particularly useful at the start of a thesis, literature review, or technical investigation when the right terminology and search routes are not yet clear. A systematic review still needs documented databases, inclusion criteria, exclusion reasons, and a reproducible search log.

How the graph works

Connected Papers says its data is connected to the Semantic Scholar Paper Corpus, licensed under ODC-BY. For each graph, it analyzes a large candidate set and presents a smaller selection with strong connections to the origin paper. Distance represents similarity, not necessarily a direct citation: the provider describes a metric based on co-citation and bibliographic coupling. Node size and color are useful orientation signals, but they are not judgments of research quality, causality, or current relevance.

A practical research workflow

1. Choose a genuinely central seed paper and record its identifier and bibliographic details.
2. Read the graph as a hypothesis map: mark clusters, older foundational works, and unexpected neighbors.
3. Use Prior Works to look for predecessors and Derivative Works to find reviews or later state-of-the-art papers.
4. Verify title, abstract, venue, year, DOI, and citation context in the original source or a second database.
5. Store accepted papers in the team’s reference manager with the search date, selection reason, and exclusion reason.

This places Connected Papers between discovery and screening. A central node is a prompt for verification, not proof that the work is relevant, rigorous, or up to date.

Operations and handover

Saved Papers and Graph History support recurring investigations. The product also lists Multi-origin graphs and views for Similar Papers, Prior Works, and Derivative Works. A team should move checked results into its existing reference manager or review log rather than treating the graph as the archive. Group plans provide administrator-controlled seats, according to the provider; Connected Papers is still not a complete systematic-review platform or library-management system.

Quality control and limits

Before citing a result, check the title, abstract, full-text version, DOI, publication venue, and the context in which a work is cited. Coverage can differ between fields, and very new, local, or poorly indexed work may not appear. Similarity can also connect papers with different methods, populations, or research questions. For reproducible reviews, use the graph as one supplementary discovery step alongside multiple databases and a predefined screening protocol.

Privacy and governance

The official privacy policy names account details, search queries, saved papers, page views, and third-party analytics. It says data is stored in Microsoft Azure and that specialized providers support authentication, email, analytics, storage, and data processing. Teams should therefore keep confidential manuscripts, unpublished project names, and personal notes out of search fields until they have checked purpose, retention, deletion, subprocessors, and whether a DPA is available. Use individual accounts, avoid password sharing, and document whether Academic or Business terms apply.

Pricing and operating cost

The official pricing page describes five Connected Papers graphs per month on the Free plan, with the listed core features included. Academic and Business plans provide unlimited graphs; Academic is intended for academic, nonprofit, or personal use, while Business is for commercial work. Group plans manage seats, and automatic renewal and unused-seat rules belong in procurement review. Prices and terms can change, so check the pricing page before buying. The wider cost also includes screening time, full-text access, reference-management software, and expert review.

Editorial Assessment

We recommend Connected Papers for a bounded discovery phase: teams with one or more good seed papers can quickly map predecessors, successors, and thematic neighbors. It creates value when every candidate is checked against the original source and recorded against explicit criteria. For systematic reviews, very new niches, or tightly controlled bibliometrics, a broader reproducible database strategy is the better foundation; Connected Papers should supplement that process rather than define it.