
Elicit
4.4
Elicit is a research assistant over an index of roughly 138 million academic papers. It finds relevant work, summarises it and — the part that saves real time — extracts specified fields from each paper into a table, so a literature review becomes a structured dataset rather than a reading list.
- Commercial use OK
- Free tier

- Pricing
- $49/mo
- Free tier
- Free tier
- Commercial use
- Permitted
- Japanese UI
- Not confirmed
- Enterprise contract
- Not confirmed
What it is good for
- Literature reviews that need to be systematic rather than impressionistic
- Extracting method, sample size and outcome across dozens of papers
- Checking whether a claim has support before repeating it
- Due diligence on a technical or scientific assertion
Strengths and trade-offs
Strengths
- Extraction into tables, not just summaries
- Every claim traces back to a specific paper
- Coverage is large enough for most non-specialist questions
- Free tier covers occasional use
Trade-offs
- Paywalled papers limit what it can read in full
- Summaries still need checking against the source
- Priced for researchers rather than for occasional business use
Editors' review
Turns a literature review into a table instead of a reading list
The distinguishing feature is structured extraction. Asking for the sample size, method and effect across forty papers and getting a table back is a different activity from reading forty abstracts, and it is where the hours actually go in a review.
Citations are first-class, so every row points at a paper. For any use where the output has to survive challenge — a claim in a report, a technical due-diligence note — that traceability is the point.
It is bounded by access and by trust. Paywalled full texts limit depth, and summaries are a starting point to verify rather than a result to quote. Used that way it is a serious time-saver; used as an oracle it is a liability.
Verdict
The best tool here for evidence-based work. Verify the rows that matter against the original papers.
Compiled from official documentation and public sources, with our own editorial judgement. We do not publish figures we cannot source.