What Elicit is
Elicit is an AI research assistant built around academic literature. It can search a large paper corpus, summarize results, answer questions with sources, chat with available full text, extract fields into tables, produce research reports, create alerts, and support screening and extraction for systematic reviews.
This is a more structured workflow than pasting a research question into a general chatbot. The value comes from keeping papers, fields, filters, and outputs together. It does not remove the need to define the question, decide which databases count, read original studies, or defend the review method.
Search, reports, and evidence tables
Begin with a precise question and explicit eligibility rules. Record population, intervention or exposure, comparator, outcomes, study design, dates, languages, publication types, and known exclusions. Run benchmark searches for papers the team already expects to find.
Elicit can summarize many results and add extraction columns. Check what text each answer used: title, abstract, available full text, figure, or uploaded document. An apparently blank field may mean the study did not report it, the text was unavailable, or the model failed to find it. Those outcomes require different treatment.
Research Reports and the Research Agent can organize a broad question into a sourced output. Treat the report as a discovery and planning artifact, not a publication-ready literature review. Verify every important statement in the cited paper, and look for missing negative results, non-English work, recent publications, corrections, and retractions.
Systematic-review workflow
Elicit's dedicated workflow can help teams import or discover records, screen papers, and extract structured data at larger scale on relevant plans. Current pricing describes Pro screening up to 5,000 papers and Enterprise scale up to 40,000, with different extraction-column limits.
Capacity is not methodological validity. Preserve the full query, search date, databases, imported files, deduplication rules, inclusion and exclusion criteria, reviewer decisions, conflicts, exclusion reasons, extracted values, and exports. Sample excluded records as well as included ones. High-impact fields and final inclusion should receive independent human review.
Do not infer PRISMA compliance from the presence of a workflow. The review team remains responsible for the registered protocol where applicable, source coverage, dual screening, risk-of-bias method, synthesis, reporting, and update process.
Exports, Zotero, and API
Basic includes Zotero import. Plus currently adds RIS, CSV, BIB, PDF, and DOCX exports, while Pro adds API access and higher extraction and review capacity. Export early in a pilot and confirm stable identifiers, abstracts, decisions, notes, custom columns, encoding, references, and attached source relationships survive.
The API can integrate Elicit output with an internal workflow, but official API terms place responsibility for third-party licenses, privacy, output protection, and validation on the customer. Store request versions and source identifiers. Generated output is provided without an accuracy or reliability warranty and should never update an evidence system without review.
Current pricing
Elicit pricing currently lists Basic as free. Plus is displayed at $11 per user per month on annual billing of $132 and adds exports, more extraction columns, clinical-trial search, and standard Research Agent, report, and systematic-review usage.
Pro is displayed at $39 per user per month with an annual billed amount shown on the page, and adds five-times usage, a dedicated systematic-review workflow, larger screening and extraction capacity, research alerts, custom extraction, explanations, templates, and API access. Scale is displayed at $89 per user per month annually, with higher usage, figure extraction, real-time collaboration, administration, and larger reports. Enterprise is custom.
Do not price only the seat. Model Agent runs, reports, screened records, extraction columns, uploaded papers, alerts, API traffic, collaboration, and human verification. Save the checkout terms because effective monthly figures, annual billed totals, and capabilities can change.
Privacy, training, and sensitive research
Elicit publishes separate terms for individual, team, authorized institutional, enterprise, API, privacy, and data-processing contexts. The DPA identifies subprocessors and points to a trust center. The relevant agreement depends on the account, not merely the product name.
Current pricing states “No training on your data by default” under Enterprise. That is useful and specific, but it should not be rewritten as a universal statement for Basic, Plus, Pro, or Scale without their controlling terms. The public individual privacy policy and terms are older than several current product and enterprise documents, so procurement should ask Elicit to resolve any material ambiguity.
Do not upload unpublished manuscripts, participant-level data, peer-review files, confidential client research, or licensed full text during an informal trial. Review model providers, subprocessors, retention, deletion, support access, data location, single-tenancy, DPA, and institutional requirements first.
Who should choose Elicit?
Elicit is a strong shortlist for researchers who need a structured path from question to papers, tables, reports, screening, and export. It is particularly relevant when the team wants AI assistance inside a visible evidence workflow instead of a standalone answer.
Choose Consensus for question-led research synthesis and Deep reviews. Choose Scite for citation context, reference checking, and supporting or contrasting statements. Choose SciSpace for a broader research suite with Agent, PDF, extraction, and writing tools. The AI research and data analysis category provides a common recall, verification, privacy, and cost checklist.
Visit Elicit
