Elicit

Elicit is an AI-powered research assistant designed to help researchers, students, academics, analysts, and evidence-focused teams find, evaluate, organize, and synthesize scientific literature. The platform was created to reduce the time and effort involved in research workflows, particularly literature reviews and evidence synthesis. Rather than operating primarily as a general-purpose chatbot, Elicit searches a large academic-paper database, extracts information from research papers, organizes findings into structured tables, and generates reports with citations that can be traced back to the underlying sources. Elicit currently provides access to a research corpus containing more than 138 million academic papers and also supports clinical-trial research.

Elicit began as a project at the nonprofit research organization Ought, where its first version of the literature-review workflow launched on August 31, 2021. In September 2023, Ought spun Elicit out as an independent public benefit corporation after transferring its intellectual property and much of its staff to the new organization. Elicit subsequently expanded from basic paper search into a broader evidence-synthesis platform with research reports, systematic reviews, research agents, structured extraction, and enterprise workflows.

The central experience starts with Find Papers, Elicit’s academic search capability. Users can enter natural-language research questions instead of relying exclusively on traditional keyword searches. Elicit uses semantic search to find papers whose meaning is relevant to the query even when they do not contain exactly the same wording. It can also use keyword-based searching and apply filters such as publication year, study type, journal, and other research attributes. This makes the system useful when the researcher has a question in mind but is not sure which terminology the literature uses.

One of Elicit’s strongest features is its ability to work with structured research tables. Search results can be displayed in rows representing individual studies and columns representing information extracted from those papers. Researchers can examine fields such as research question, methodology, population, intervention, outcome, study type, or other custom criteria. Additional columns can be added to investigate particular aspects of a literature set, transforming a simple paper search into a structured evidence-review workflow.

Elicit also provides Research Reports, which automatically organize information from multiple research papers into a structured report. The platform searches for relevant studies, extracts key information, and synthesizes the findings into a readable output. Elicit emphasizes that AI-generated claims in its research workflows are supported by sentence-level citations to the underlying sources, allowing users to inspect where conclusions came from instead of receiving uncited summaries.

For deeper research, Elicit offers a Research Agent. This agent is designed for more complex, multi-step research tasks rather than a single search. It can gather evidence across academic databases, journals, public data sources, and documents uploaded by the user. The current product positioning focuses on high-stakes research and decision-making, with the agent able to combine searching, analysis, evidence extraction, and synthesis into longer workflows.

Elicit’s Systematic Review workflow is aimed at researchers conducting structured evidence reviews. The platform has introduced a dedicated workflow aligned with PRISMA 2020, with screening and extraction steps that are designed to be reproducible and auditable. On the current Pro plan, the dedicated systematic-review workflow can screen up to 5,000 papers, while higher plans provide substantially greater scale. Enterprise capabilities extend this to as many as 40,000 papers.

A major benefit of this workflow is automated screening and extraction. Researchers can define screening criteria, have Elicit apply those criteria consistently across candidate papers, and extract selected information into a structured table. Elicit’s API documentation also exposes programmatic endpoints for search, systematic reviews, screening, extraction, and report generation, allowing organizations to incorporate parts of the research workflow into their own applications.

Elicit can also work with full-text research papers. The platform provides Paper Chat and full-text access where available, allowing users to ask questions about a paper and inspect details beyond the abstract. This is valuable for researchers who need to examine methods, results, limitations, or other portions of a study that may not be represented accurately by metadata alone. The Basic plan includes unlimited chat with papers when full text is available.

The platform supports custom extraction and increasingly sophisticated analysis of research data. Users can define the information they want to collect from papers instead of relying only on preset fields. Elicit can extract details from large collections of research papers and, on higher plans, can interpret information contained in figures. This makes the system useful for reviews in fields where important evidence may occur inside tables, figures, or other structured parts of a scientific paper.

Elicit also provides Research Alerts that allow users to monitor new research related to their interests. Pro and higher plans can subscribe to personalized research alerts, helping researchers stay aware of new publications without repeating the same search manually. This can be particularly helpful for rapidly changing scientific fields where new literature appears frequently.

The platform emphasizes evidence traceability as one of its defining characteristics. Elicit states that it supports AI-generated claims with sentence-level citations from underlying research sources. It also warns users that Elicit is an assistant rather than a replacement for researcher judgment, and recommends reviewing the original papers to confirm summaries and extracted information. This is important because AI can summarize a study accurately in some cases while still missing context, limitations, or methodological nuance.

Elicit provides tools for paper organization and data export. Researchers can save papers to their Elicit Library, organize research material, and export reports or research tables. Depending on the plan and workflow, exports are available in formats including PDF, Word, CSV, Excel, RIS, and BIB. RIS and BIB exports can be imported into reference-management software such as Zotero, Mendeley, and EndNote, allowing Elicit to operate alongside existing academic reference workflows instead of requiring researchers to abandon them.

The platform also offers clinical-trial search. Current Plus and higher plans provide unlimited searching across more than 500,000 clinical trials, while Elicit’s API can search ClinicalTrials.gov as a dedicated corpus. This strengthens the product’s usefulness for healthcare research, medical evidence synthesis, clinical development, and related research applications.

Elicit’s current positioning extends beyond universities. The company explicitly markets the platform to pharmaceutical companies, medical-device and technology companies, policy and government organizations, consumer-goods companies, industrial companies, and software and technology organizations. The platform highlights applications such as preclinical research, clinical development, medical affairs, marketing and sales research, business development, product research, technology research, and evidence-backed policy development.

The platform’s API and MCP support significantly expands its usefulness for technical teams. The Elicit API provides programmatic access to semantic search over more than 138 million papers and to automated research reports. The API currently supports research-paper search, clinical-trial search, report generation, and systematic-review workflows. Elicit’s MCP server exposes the same capabilities to compatible AI clients, including Claude Desktop and Claude Code, allowing researchers and developers to bring Elicit-powered academic research into other AI workflows.

Elicit also provides examples of integration into automated research systems. Its 2026 API announcement describes use cases including agents that evaluate scientific evidence, healthcare research systems, and coding agents that find supporting papers for claims in a draft. This demonstrates a shift from Elicit being simply a research website to becoming an infrastructure layer for evidence-backed AI systems.

The platform’s pricing follows a freemium subscription model. The current Basic plan is free and provides unlimited paper searching across more than 138 million papers, unlimited summaries, full-text paper chat where available, source viewing, Zotero import, and limited access to Research Agent and Research Reports. The Plus plan costs $11 per user/month when billed annually, while Pro costs $39 per user/month annually and Scale costs $89 per user/month annually. Monthly billing is higher: the current pricing page lists Pro at $49 and Scale at $169 when billed monthly. Enterprise plans use custom pricing.

The Plus plan is positioned for deeper research and adds exports, clinical-trial searching, additional table columns, and standard usage for Research Agent, Research Reports, and systematic literature reviews. Pro is designed for systematic reviews and adds five-times the standard usage for major AI research features, a dedicated systematic-review workflow, 5,000-paper screening, custom extractions, research alerts, explanations for AI-generated answers, and API access. Scale increases usage further and adds live collaboration, figure interpretation, an admin panel, and larger report data-source limits. Enterprise adds organizational controls, custom data sources, SSO/SAML, user analytics, custom deployments, and unlimited search API access.

Elicit is not primarily a writing assistant in the same sense as Grammarly, Jenni AI, or Paperpal. Its greatest strength is evidence discovery and synthesis, although it can generate reports, explanations, summaries, and other research outputs. It is therefore particularly useful when the user’s priority is finding what the scientific literature says and organizing evidence rather than simply improving grammar or generating prose.

Independent review information should be treated carefully. There are multiple G2 profiles with the name “Elicit,” including an unrelated marketing-search product that has a 4.5/5 rating from one review. The current G2 profile for the Elicit AI research product shows 0 reviews, so there is currently no reliable G2 star rating for the academic research platform. This distinction is important because using the older 4.5/5 G2 figure would incorrectly attribute reviews belonging to a different product.

Elicit is best suited to research-intensive users and organizations that need to find scientific literature, compare studies, conduct literature or systematic reviews, extract structured evidence, and build evidence-backed reports. Its combination of a large academic corpus, research-agent workflows, systematic-review capabilities, source-linked outputs, reference exports, API access, and MCP connectivity makes it one of the more research-specialized AI platforms available.

Pricing

• Basic: Free.
• Plus: $11/user/month when billed annually ($132/year).
• Pro: $39/user/month when billed annually ($468/year).
• Scale: $89/user/month when billed annually ($1,068/year).
• Enterprise: Custom pricing.

Monthly pricing is higher: Pro is $49/month and Scale is $169/month.

Review

The current G2 profile for the academic Elicit product has 0 reviews, so there is no reliable G2 rating to report. An older G2 profile named “elicit” with a 4.5/5 rating belongs to a different marketing-search product and should not be used for Elicit AI.

Elicit’s strongest documented value is its research workflow: large-scale academic search, structured extraction, systematic reviews, research reports, citations, and evidence synthesis. The platform itself emphasizes that AI results should be checked against the original literature.