Use Elicit to find hundreds of relevant papers

Elicit is positioned as a fast way to generate and manage an “insane number” of relevant academic papers—so researchers can avoid missing key studies when building a literature base. The core workflow starts with running a query, reviewing the top results, and repeatedly clicking “load more” to expand the list until the search reaches diminishing returns. In a live example using sperm-count research, a single query like “sperm count measures” quickly produces dozens of papers, and the user can keep extending the results list, delete off-topic items, then prioritize what remains.

The next lever is running multiple independent queries inside the same project. Instead of relying on one search term, Elicit supports re-framing the same topic with different filters embedded directly in the query—such as restricting to papers published after 2000 or switching from “sperm count measures” to “sperm density measurement.” Each query acts like a separate search “tab,” but all results can be analyzed together later. This matters because different phrasings and inclusion criteria surface different subsets of the literature; once one query slows down, the workflow encourages starting a new one rather than forcing more value out of the same search.

Citation searching adds a second, compounding pathway for discovery. After selecting a promising paper, Elicit can surface “like these” papers by traversing the citation graph in both directions: backward through references and forward through later citations. The ranking is tied to relevance to the original query, making it easier to expand from a small set of strong candidates into a broader set of related work. The same mechanism can be tuned for specific research needs. For example, if a user finds a meta-analysis and wants the underlying evidence, filters can be applied to pull out primary studies mentioned in that review—such as limiting to RCTs (randomized controlled trials). If the goal is to see what happened after a particular paper (e.g., whether it was challenged or extended), date filters can restrict results to later citations.

Once enough papers accumulate—72 in the example—the workflow shifts from discovery to organization. Elicit can combine papers from all queries into a single table, enabling systematic screening against inclusion and exclusion criteria. The transcript also highlights a strategy for managing the “research explosion”: generate many candidate search terms first (even while results are loading), then do a second pass to refine queries based on what actually appears. Every search attempt is saved in the notebook history, allowing researchers to return later and continue expanding.

Finally, the tool is framed as a supplement rather than a replacement for established search methods, especially for formal systematic reviews. The practical promise is coverage: Elicit helps surface papers traditional searches may miss, then supports filtering and citation-graph expansion to reach a more comprehensive literature set—without getting lost in the volume.