AI Agent Index
ByHeather MacAvelia·Independently reviewed·Published Mar 18, 2026·Updated Sep 24, 2026
Independently verified against live vendor data on Sep 24, 2026.

AI research assistant for systematic reviews, research reports and a Research Agent over 138M papers and 545K clinical trials. MCP server and API. Free plan, Pro from $49/user/mo billed annually (USD).

How we scored it

Autonomy

4/5

Integrations

4/5

Pricing clarity

4/5

Evidence

5/5

Setup

5/5

The facts

Elicit is an AI research assistant built by Elicit Research, PBC, a public benefit corporation based in Oakland, California. It searches over 138 million academic papers from Semantic Scholar, PubMed and OpenAlex, plus 545,000 clinical trials, and supports work from a quick cited answer to a full systematic literature review. Its Research Agent, launched in August 2026, gathers evidence across the scientific literature, clinical trials, patents, regulatory and biology databases, web sources and a team's own uploaded documents, and turns it into cited reports, tables, figures, analyses and slides. Research Reports, a Systematic Review workflow that supports PRISMA 2020, a shared Library and research Alerts cover the rest of the review process. Elicit says it is used by over 5 million researchers, and its customer page shows 41 organizations, including NASA, NIH, the U.S. CDC, Harvard, Stanford, Yale, Google, Johnson & Johnson, Takeda, Unilever and the Gates Foundation. Elicit publishes an MCP server and an API on the Pro, Scale and Enterprise plans, covering paper search, reports and systematic reviews, so the same capabilities can run inside Claude, Claude Code, ChatGPT, Gemini or any other MCP-compatible client. Papers can be imported from Zotero or uploaded as PDFs, RIS and BIB files, results export to RIS, CSV, BIB, PDF and DOCX, and a browser extension gives full-text access to papers through a user's own journal subscriptions. The paper corpus is drawn from Semantic Scholar, PubMed and OpenAlex. Prices are in US dollars, and the pricing page has separate Industry and Academic ladders. On Industry pricing, Basic is free, Pro is $49 per user per month billed annually ($588 a year) or $75 billed monthly, and Scale is $169 per user per month billed annually ($2,028 a year) or $279 billed monthly. Academic pricing adds a Plus plan at $11 per user per month billed annually or $19 monthly, with Pro at $39 or $69 and Scale at $89 or $149. Enterprise is priced by quote. Elicit does not offer a separate nonprofit discount and points nonprofits to its academic plans. Every plan draws on a single monthly usage pool shared across the Research Agent, Reports and Systematic Literature Reviews, and unused usage does not roll over. Basic has limited usage. The API and MCP server start on Pro, the Pro systematic review workflow screens up to 5,000 papers, figure extraction and real-time collaboration start on Scale, and SSO and SAML, single-tenant deployment and screening of up to 40,000 papers are on Enterprise. Pro and Scale users can turn on extra usage, billed in $100 increments up to a monthly limit they set, and Basic and Plus users who reach their limit keep Find Papers and Chat with Papers until the next cycle. Elicit suits researchers and research teams in pharma, medical devices, academia, policy and consumer goods who run literature reviews or need cited evidence for research decisions. Researchers who want to see where peer-reviewed studies agree on one specific question should also look at Consensus, those checking whether later papers support or contrast a finding should look at Scite.ai, and those exploring a field through visual citation maps should look at Litmaps. Current state Q3 2026: Elicit launched its Research Agent in August 2026 and moved every plan to a monthly usage pool in June 2026. In September 2026 it turned its Library into a shared research hub, where the Research Agent can search a team's collections and save papers to them. Its trust center publishes a SOC 2 Type 2 report, a 2026 bridge letter and Cyber Essentials, and names Anthropic and OpenAI among its subprocessors for AI responses, with custom models hosted on CoreWeave. Enterprise plans state no training on customer data by default, and the API terms state that API input is not used to train AI models unless a user opts in. Elicit is designed for use in English. It raised a $22M Series A led by Spark Capital and Footwork in February 2025.

Pricing

freemium · $49/mo annual

View pricing ↗

Segment

both

Setup

easy

Verified

Sep 24, 2026

Transparency

Mostly Public

Contract

Monthly or Annual

Data training

Not Disclosed

Human in loop

Optional

Capabilities

deep-researchcitationsdata-analysispdf-analysisautonomousweb-search

Pros & Limitations

Editorial assessment

Pros

  • ✓Covers research from search to a cited output: paper search, Research Reports, a Systematic Review workflow that supports PRISMA 2020, and a Research Agent that produces reports, tables, figures and slides.
  • ✓Searches over 138 million papers and 545,000 clinical trials, and the Research Agent adds patents, regulatory and biology databases, web sources and uploaded documents.
  • ✓Publishes an MCP server and an API on Pro and above, so search, reports and systematic reviews can run inside Claude, ChatGPT and other MCP clients.
  • ✓Publishes a SOC 2 Type 2 report on its trust center, and its customer page shows 41 organizations, including NASA, NIH, Harvard, Takeda and Unilever.

Limitations

  • ⚠Usage is a single monthly pool that does not roll over, and larger Research Agent sessions and reviews, figure extraction and premium PDF parsing use more of it.
  • ⚠The API and MCP server start on Pro at $49 per user per month billed annually, and SSO, SAML and single-tenant deployment are on the quote-based Enterprise plan.
  • ⚠Enterprise states no training on customer data by default, and the terms for the Basic, Plus, Pro and Scale plans carry no equivalent training statement.
  • ⚠Designed for use in English, and Elicit states that quality in other languages varies widely and that non-English use draws down usage faster.

Technical Details

Deployment
web
Model architectureAnthropic and OpenAI models (named subprocessors for AI responses) plus custom ML models hosted on CoreWeave, over a 138M-paper corpus from Semantic Scholar, PubMed and OpenAlex
Avg setup timeNo published setup duration. The free Basic plan starts with an account sign-up, and the MCP server connects through OAuth from Claude, Claude Code or another MCP client.
Autonomous rateThe Research Agent plans and runs multi-step research across its data sources and returns cited outputs without per-step approval, and the Systematic Review workflow screens papers and extracts data against the user's criteria. Elicit publishes no autonomous completion rate, and plans bill on a monthly usage pool rather than per task.
Integrations
MCP ServerAPIZotero (import)Semantic ScholarPubMedOpenAlexClinicalTrials.govBrowser extension (full-text access)RIS, CSV, BIB, PDF and DOCX export
Security
SOC 2 Type IIGDPRCCPA

Similar agents

Rating

4.4/ 5
★★★★★

Editorial score

How we score this →

Recognition

MCP Server VerifiedListed 2026
Embed these on your site →

Industries

ResearchPharmaHealthcareEducationPublic Sector

Leave a review

Never displayed publicly.

Agent Stacks

See workflow stacks that feature Elicit.

Compare

Related Content

Is this your tool?

Claim this listing free to get your embeddable badges, a review collection link, and the Claimed badge on your listing.

Claim listing →

Free · Every Two Weeks

AI Agent Price & Rating Tracker

Price changes, new agent launches, acquisitions, and rating updates across the AI agent market. Verified against live vendor data, not vendor marketing.

No spam. Unsubscribe anytime. We never share your email.