Elicit

• Semantic search & retrieval of relevant academic papers
• Automatic summarization of research findings relative to a question
• Data extraction from PDFs (tables, quantitative results)
• Guided systematic review workflows (search → screening → extraction → reporting)
• Customizable data extraction columns and table structures
• Export of extracted data and tables (CSV, RIS, BIB)
• Editable reports: you can revise criteria or steps and see updates
• Alerts / monitoring of new papers in your topic area
• Integration with reference managers and ability to upload your own PDFs
• Transparent reasoning and citations: linking AI outputs back to original sources

Elicit is an AI research assistant designed to reduce the time and effort researchers spend on literature reviews, evidence synthesis, and data extraction. Built by Ought, Elicit employs language models and structured reasoning techniques to help users find relevant academic papers, extract key data, summarize findings, and create research reports—all in an interactive, controllable workflow.

To begin, a researcher poses a question or topic; Elicit then conducts a semantic search across millions of scholarly papers to identify broadly relevant works. Instead of simply returning results, Elicit summarizes findings in relation to the research question, highlights key points, and surfaces quantitative results where possible. The extraction engine delves into PDFs, pulling out tables, figures, and numeric data into structured formats (tables with customizable columns).

In more advanced workflows, Elicit supports systematic reviews, guiding researchers through stages: searching literature, screening titles/abstracts, extracting data, and producing a final report. The user retains full control: you can revise earlier steps (e.g. change screening criteria or extraction instructions) and the system will recalculate downstream results accordingly. This flexibility ensures transparency and adaptability in complex research workflows.

Elicit’s tiers allow varying levels of usage: the Basic (free) tier offers unlimited search, summaries of a few papers, and modest data extraction. As users scale, the Plus and Pro plans provide expanded extraction allowances, export options, more table columns, and access to high-accuracy extraction modes. For teams and institutions, the Team tier enables collaborative research, shared workflows, and administrative oversight.

A strength of Elicit lies in its emphasis on linking AI outputs back to source literature (citation transparency) rather than providing unsupported assertions. This helps mitigate hallucination risk and enables users to verify claims easily. Although not perfect—its outputs are typically around 80–90 % accurate, requiring human oversight—it substantially accelerates literature work. 

Researchers in academia, policy, health economics, medical research, and other domains use Elicit to scale evidence reviews, explore new hypotheses, and stay current in their fields. As the volume of published literature continues to grow exponentially, tools like Elicit become essential for maintaining rigor and speed in scientific discovery.