Elicit

Elicit is an AI research assistant built specifically for academic literature, made by Elicit (spun out of the nonprofit research lab Ought in 2023). Instead of searching the open web, it queries a corpus of 125+ million papers via Semantic Scholar, so answers come grounded in actual studies rather than blog posts. Ask a research question and Elicit returns a ranked list of relevant papers with a one-sentence, citation-backed summary of what each found.

Its signature feature is data extraction. Select a set of papers and Elicit builds a comparison table, pulling structured fields out of each PDF — sample size, population, intervention, methodology, outcomes, limitations — the tedious core of any systematic review. Reviewers report cutting screening and extraction time dramatically, which is why the tool is popular in evidence synthesis, biomedicine, and policy research. Competitors overlap only partially: Consensus focuses on yes/no answers across studies, ResearchRabbit on citation mapping, and Scite on whether citations support or dispute a claim; none automate extraction tables the way Elicit does.

The Systematic Review workflow goes further, letting you define inclusion and exclusion criteria and having the AI screen hundreds of abstracts against them, flagging its confidence for each.

Key Features

• Semantic search across 125M+ papers from the Semantic Scholar corpus
• Automated data extraction from PDFs into structured comparison tables
• Screening workflow that applies your inclusion/exclusion criteria to abstracts
• One-line, citation-linked summaries of each paper’s findings
• Upload your own PDFs and run the same extraction on them
• Export tables to CSV or BibTeX for downstream analysis
• Notebook-style follow-up questions scoped to your selected papers

Pricing

• Basic (Free): Limited monthly credits for search, summaries, and small extractions
• Plus (~$12/month): More credits, larger extractions, full-text search on more papers
• Pro (~$49/month): High limits aimed at active systematic reviewers
• Team/Enterprise: Shared workspaces and custom terms

Best For

Graduate students and academics running literature reviews, medical and public-health researchers doing evidence synthesis, and analysts who need structured comparisons across dozens of empirical studies rather than a chatbot’s paraphrase.

Limitations

Extraction accuracy is good but not trustworthy enough to skip verification — you must spot-check against source PDFs. Coverage is strongest in biomedical and quantitative fields; humanities and non-English literature are sparse. Heavy review work burns through credits fast, pushing serious users to the Pro tier.