How to Search for Sources and Manage Them

Inciteful graph

The Open Citations movement has facilitated access to bibliographic metadata. Combined with the advances in artificial intelligence, it has enabled the creation of new search tools.

Literature mapping tools help researchers find articles by exploring connections between publications. Most use one or more 'seed papers' as a starting point. They often use citations (articles citing or cited in the seed papers), or articles that appear in the bibliographies of several papers), algorithms and artificial intelligence. The results are displayed as a map, with the closest articles grouped together.

	Freemium Uses co-citation and bibliographic coupling to find articles that have highly overlapping citations and references 5 graphs per month with a free account Uses the Semantic Scholar database
	Free No account creation required. Builds a citation network graph from one or several seed paper(s), or shows the connections between two papers There is a Zotero plugin allowing to create graphs directly from an item of your Zotero library
	Freemium Enter keywords, the DOI or the name of an author, and find articles linked to it through citations
	Open source and free No cookies, no user-tracking Allows to build a citation network, starting from a source article
	Non-profit and open source Ask a research question and the results will be presented on a map with the 100 most relevant documents grouped by sub-topics
	Free (users can make donations) Does not sell users' data to third parties Rather sophisticated, it encourages jumping from paper to paper without getting lost in the rabbit hole You can create collections of papers inside Research Rabbit and synchronise them with your Zotero library; you can also import a Zotero collection into Research Rabbit for further exploration Allows to explore the network of researchers, to understand who publishes with whom

Cautious: AI tools "hallucinate" (give  false answers that seem credible, especially if you are unfamiliar with the subject) and even make up bibliographic references. They can also be "infected" to spread false narratives.

Here is a selection of tools designed specifically for an academic use.

	Freemium Enter a search question Generates a synthesis from the papers, and also provides a "consensus-meter" (proportion of papers answering yes or no to the search question). Zotero integration
	Freemium; account creation required Enter a search question Gives a one-sentence abstract summaries, extracts details from papers into an organised table Zotero integration
	Freemium; based in a European country (Norway) Starts with a PDF file; Keenious analyses it to find relevant papers Integrated with Google docs and Microsoft Word; you can use your own drafts to search for new papers
	Free and AI-powered Looks like Google Scholar, gives normally less but more relevant results Explores the citation graph both ways (citing and cited papers), whereas Google Scholar goes only one way (citing papers). Distinguishes between "highly influential citations" and "background citations" A project at the Allen Institute for AI, a non-profit research institute created by Paul Allen, Microsoft co-founder
	Freemium, you need to create an account Enter a search question with keywords Helps readers to understand articles more easily; provides a synthesis of top 5 articles You can ask follow-up questions to the "copilot", and even ask it to explain tables or figures You can also upload a PDF and ask Scispace to explain the content of the text There is a Chrome add-on
	Paid Uses Chat GPT for text generation Enter a natural language question Generates a summary with citation; hovering over the citation allows you to see the part of the text that generated the passage

Deep research tools promise better, more comprehensive reports and literature reviews, but they "think" slower, taking minutes rather than seconds to answer. They usually ask you a few questions to better define the purpose of your research. It is an iterative process, they run multiple searches and refine the search by taking into account the results of previous searches. Users can see the different steps of their report generation, and intervene if necessary. Be careful! They can also hallucinate and often go off topic. The sources used are not always academic, and usually limited to open-access articles.
Deep research can be very expensive. Here are some affordable options (and two less affordable) to try. Cross-checking required!

	Open source, a project of the Allen Institute for AI and Semantic Scholar Searches for academic sources in Semantic Scholar Beta version, a limited corpus of articles.
	Basic search accessible without account, account required for enhanced research. 3 free enhanced queries each day. Seems to use only a minority of the sources in the report and often goes off topic.
	You need a 70$ pro month subscription (in March 2025) to access this service Uses academic sources Asks you follow-up questions to improve your query Iterative process with refinement of the queries Provides follow-up questions that help continue the research process
	Open source, research prototype Writes a "Wikipedia-like report" or a roundtable conversation on your search topic English only Uses a non-academic search engine (Bing)
	Freemium (5 searches per-month with the free plan) Uses Semantic Scholar, usually title and abstract only Invites you to have a conversation with it, as you would do with a colleague asks you a few questions to clarify your search