Elicit

Elicit is an AI research assistant built for working with academic papers. You ask a research question, and it searches a database of well over a hundred million papers, summarises what it finds and can extract structured data from many papers into a table. It also supports systematic review workflows, which is why it is popular with students, academics and evidence-based teams.

Quick facts

• Made by: Elicit, a research-focused AI company
• Free plan: Yes. Basic includes unlimited search, summaries and chat with papers, but only 2 automated reports a month.
• Paid plans: Pro and Scale per user, plus custom Enterprise. Older pages also list Plus and Team plans.
• Works on: Web, with an API launched in 2026
• Pricing last checked: October 2026, using several recent published sources. Elicit restructured its plans in 2025 and 2026, so older pages differ. Confirm on Elicit’s official pricing page.

What Elicit is best at

• Finding relevant papers: search by concept, not just keywords, across a very large paper database.
• Extracting data into tables: pull study design, sample size, results and limitations from many papers at once.
• Systematic reviews: screening and extraction workflows that follow a structured review process.
• Automated reports: synthesised write-ups of what the literature says on a question.
• Working with your own library: import papers from Zotero and ask questions about full texts.

Plans and pricing

Older pages list a cheaper Plus plan at about $10 to $12 a month and a Team plan at about $79, which appear to predate the restructure, so treat them with care. Limits such as report counts and table columns differ between sources, so confirm the exact allowances on Elicit’s official page.

Best for

• Graduate students and researchers doing literature reviews.
• Analysts and policy teams who need evidence tables from many studies.
• Anyone who wants answers grounded in papers rather than general web content.

Less suited to

• General questions that need current news or web sources. Perplexity covers the wider web.
• Casual users, because the jump from the free plan to Pro is large.
• Work that depends on research outside academic papers, such as internal documents.

How to get started

1. Create a free account and ask a specific, answerable research question.
2. Read the sources behind each summary rather than relying on the summary alone.
3. Try extracting a few fields, such as sample size and main result, from a small set of papers.
4. Check the extracted values against the papers, because extraction can contain errors.
5. Upgrade only if you need systematic review screening or many reports.

Prompt ideas to copy

What is the effect of [intervention] on [outcome] in [population]? Find randomised trials and summarise the findings in a table.

Extract the sample size, study design, main result and limitations from each of these papers.

Find papers that disagree with [claim] and summarise why they differ.

Screen these abstracts against my criteria: [inclusion and exclusion criteria]. List which are included, excluded or unsure.

Limitations and things to watch

• AI summaries and extractions can be wrong, so verify important claims in the original papers.
• Free reports and table columns are very limited.
• Coverage and quality depend on the underlying paper database.
• It supports research but does not replace expert judgement or formal review methods.

Alternatives

Compare NotebookLM for questions about your own sources, Perplexity for cited web research, and Claude or ChatGPT for reading and discussing papers you upload.