Elicit

What is Elicit?

AI research assistant for systematic literature review. Searches a corpus of ~125 million academic papers, extracts structured data from each into a comparison table, and summarises findings across studies.

Main features

• Semantic paper search across ~125M papers
• automatic data extraction into custom table columns
• systematic review workflow
• screening with custom inclusion criteria
• Elicit Reports
• PDF upload and extraction
• citation export
• notebook for multi-step research.

Practical use cases

• Running a systematic literature review
• screening hundreds of abstracts against inclusion criteria
• extracting sample sizes, methods and outcomes into a structured table
• finding papers on a concept rather than a keyword.

Who is it for?

• Academic researchers, PhD students, systematic reviewers, evidence-synthesis teams, pharma and biotech research, policy analysts.

Pricing, free access & limits

Freemium with credit-based extraction; per-seat team and enterprise plans

Plans described in the source

Basic free with a monthly credit allowance. Plus around $12/mo billed annually (about $15 monthly) with a larger credit pool and unlimited search. Pro around $49/mo annually for high-volume extraction and reports. Team and Enterprise priced per seat on request. Credits are consumed per paper column extracted.

Usage limits to consider

Credit-metered: each cell of extracted data from each paper consumes credits, so a 200-paper, 8-column extraction is expensive. Free tier credits are quickly exhausted by any real review.

Strengths & tradeoffs

These points summarize the supplied research, rather than hands-on test results.

Strengths noted in the source

• Purpose-built for systematic review rather than generic chat
• extraction into structured tables is a real time-saver
• grounded strictly in the paper corpus so hallucination risk is lower
• transparent about which paper each claim came from.

Limitations to consider

• Corpus skews heavily toward English-language and open-access literature
• paywalled full texts often unavailable so extraction works from abstracts
• credit pricing makes large reviews costly
• extraction accuracy still requires human verification for publication-grade work.

Platforms, languages & integrations

Platforms

• Web (browser only)

Languages

• Interface English; indexed literature is predominantly English

API access

Enterprise API available on request

Integrations and exports

• Zotero, BibTeX/RIS citation export, CSV export, PDF upload

Evaluation notes

Always spot-check extracted values against the source PDF - accuracy is good but not publication-grade unverified. Budget credits against the number of papers multiplied by the number of columns, not by seats.

Explore similar tools

Other tools in AI Research Tools. Compare their specific features and limits for your task.

Frequently asked questions

Sources & editorial notes

Based on the directory editor’s supplied research. The source does not provide a verification date for this row. Importing a record does not independently verify every product claim.

• Official product website
• Official pricing