Elicit

What is Elicit?

Elicit is an AI research assistant developed by Elicit Research PBC. It is built for literature discovery and evidence synthesis rather than casual chat or open-ended content generation. A researcher can start with a natural-language question, search for relevant papers or clinical trials, inspect source material, and organize findings into tables, reports, or a research project.

The product is especially useful when the answer needs to be traceable to evidence. Elicit shows the sources behind its outputs and supports workflows such as paper screening, data extraction, systematic reviews, and research reporting. It can also work with uploaded research materials and, in Research Agent workflows, broader sources such as public or web data.

How researchers use Elicit

A typical workflow begins with a research question rather than a keyword list. Elicit retrieves potentially relevant papers, summarizes them, and helps the user compare study characteristics or extract information into structured columns. The researcher can then review the underlying sources, refine the question, screen papers, and export the resulting evidence base.

For a more open-ended investigation, Research Agent can turn a question into a research session that produces reports, tables, analyses, figures, visualizations, slides, or documents. These outputs are useful starting points, but they still require subject-matter review. Elicit is an aid for organizing and synthesizing research, not a replacement for reading critical papers or making expert judgments.

Core capabilities

Literature and clinical-trial search

Elicit searches a large academic-paper database and can also search and summarize clinical-trial records. Its natural-language approach is intended to help users find relevant studies even when they do not already know the exact terminology used in the literature.

Paper Chat and evidence extraction

Users can ask questions about papers with full-text access, view supporting sources, summarize studies, and extract custom fields. This is particularly useful for building comparable evidence tables containing study designs, populations, interventions, outcomes, or other details selected by the researcher.

Systematic review workflows

Elicit includes workflows for screening papers and extracting study information for systematic literature reviews. Paid tiers differ in review capacity and related usage limits, so the practical fit depends on the size of the review and how much automation is needed.

Research reports and Research Agent

Research Reports and Research Agent extend Elicit beyond individual paper summaries. They can organize findings into a cited overview and work across academic literature, clinical trials, selected public or web sources, and uploaded internal materials. The resulting reports and visual artifacts can be inspected and exported rather than treated as unquestionable conclusions.

Projects, collaboration, and exports

Libraries, projects, alerts, and shared research sessions help users maintain an ongoing evidence workflow. Exports can include formats such as PDF, DOCX, CSV, RIS, and BIB. Zotero import is available, while collaboration and administrative controls depend on the plan.

What distinguishes Elicit from general AI search

Unlike a general-purpose chatbot, Elicit is organized around research objects such as papers, trials, extracted fields, evidence tables, and citations. It is closer to a research workspace than a conversational answer engine. Readers comparing it with an AI search tool such as Perplexity should focus on workflow: Elicit emphasizes literature screening, structured extraction, and review processes, whereas general AI search products are typically optimized for fast web-oriented answers.

Elicit also differs from a standard document chatbot. Its purpose is not only to answer questions about one uploaded file, but to help locate relevant studies, compare multiple sources, create structured evidence, and assemble a research deliverable. It can complement general assistants such as Claude when a project requires a dedicated evidence-gathering workflow, although Elicit's underlying model providers are not fully disclosed.

Who is Elicit for?

Elicit is a strong fit for academic researchers, students, scientific teams, clinical and medical-affairs groups, consultants, analysts, patent researchers, and organizations that need documented evidence for a decision. It is most useful when the work involves many papers, repeated screening or extraction tasks, or a report that must remain connected to its sources.

It is less suitable for users seeking a general conversational assistant, creative writing tool, coding environment, meeting transcription service, or image-generation product. Someone who only needs a quick factual answer may find Elicit's research-oriented workflow more involved than necessary.

Pricing and access

Elicit has a free Basic plan with unlimited academic-paper search, paper summaries, full-text paper chat, source viewing, and Zotero import. Advanced features are usage-limited on the free tier, particularly Research Agent and Research Reports.

The pricing page lists Plus from $11 per user per month when billed annually, Pro at $39 per user per month billed annually, and Scale at $89 per user per month billed annually. Enterprise pricing is custom. Higher plans increase usage and capacity for research workflows and may add collaboration, administrative controls, larger systematic-review workloads, API access, and enterprise security features. Pricing can vary by billing mode or customer type, and academic pricing may be available through the pricing selector.

The Elicit API and MCP server are available to eligible paid customers, including Pro, Scale, and Enterprise users. MCP support can connect Elicit research capabilities with compatible clients such as Claude Desktop or Claude Code, while the browser application remains the main way most users interact with the product.

Limitations and privacy considerations

Elicit is specialized, and its strongest results depend on a clearly framed question and careful source review. AI-generated summaries, extracted fields, and reports can contain errors or omit important context, so high-stakes medical, scientific, legal, or business conclusions should be verified against the original material.

English is the primary supported language, and Elicit says its search, summaries, and extraction are optimized for English-language papers. Some workflows may handle other languages, but quality can vary. Advanced usage, large-scale review capacity, collaboration, API access, and enterprise controls are restricted by plan. No official native mobile or desktop application was verified.

Privacy also depends on the plan and agreement. Elicit processes account information, usage information, and submitted content to provide the service. Enterprise materials state that customer data is not used for training by default, while general privacy and terms materials allow processing related to service provision and improvement. Organizations handling sensitive research should review the applicable privacy policy, data-processing terms, retention provisions, and enterprise documentation before uploading material.

Bottom line

Elicit is best understood as an evidence-synthesis workspace with AI-assisted search, screening, extraction, and reporting. Its advantage is not general conversational breadth; it is the way research questions, papers, structured evidence, citations, and exports fit together. For users conducting literature reviews or evidence-based analysis, it can reduce repetitive organization work. For casual questions or unrestricted general assistance, a broader AI assistant may be a better fit.