AI Search

Consensus: AI search across 220M scientific articles with citations

Consensus

  • Free: 15 Pro requests + 3 Deep Searches/month; Pro - $10/month
  • Web, API, ChatGPT (Custom GPT)
  • 220M+ peer-reviewed articles in the database
  • Scholar Agent on GPT-5 - complex tasks in minutes
  • 170+ university libraries offer Pro for free
10M+
Active users
220M+
Articles in the database
$40M
Series A, May 2026
2021
Year founded

Scientific search without hallucinations - only peer-reviewed sources

Try for free
  • Free: 15 Pro requests + 3 Deep Searches/month; Pro - $10/month
  • Web, API, ChatGPT (Custom GPT)
  • 220M+ peer-reviewed articles in the database
  • Scholar Agent on GPT-5 - complex tasks in minutes
  • 170+ university libraries offer Pro for free
10M+
Active users
220M+
Articles in the database
$40M
Series A, May 2026
2021
Year founded

Scientific search without hallucinations - only peer-reviewed sources

Consensus is a specialized AI search engine for peer-reviewed scientific literature, launched in 2022 by a San Francisco-based company. Unlike ChatGPT or Perplexity, the service does not generate answers from training data: the AI is engaged after searching real articles, which physically precludes fictitious references.

Key Fact: In May 2026, the company closed a Series A round of $40 million from Union Square Ventures, GreatPoint Ventures, Nat Friedman, and Daniel Gross - confirming its leadership position in the AI scientific literature search market.

Key Features of Consensus

  • Consensus Meter - a visual "balance of evidence" scale: for binary questions ("Does zinc help with colds?") the system shows how many peer-reviewed articles lean toward "Yes," "No," or "Maybe," with direct links to each source.
  • Pro Search - analyzes up to 20 relevant peer-reviewed articles and synthesizes a narrative response with inline citations; filters by design type (meta-analysis, RCT, systematic review), journal quartile Q1-Q4, and citation threshold.
  • Deep Search - scans up to 1,000 articles in 2-5 minutes, identifies contradictions and gaps in the literature, and creates a structured report with a research timeline and a clickable bibliography.
  • Scholar Agent - a multi-agent system based on GPT-5 and OpenAI Responses API (2025): four specialized agents (Planning, Search, Reading, Analysis) automatically coordinate complex research tasks.
  • Medical Mode - narrows the search to ~8 million articles from the top 1,000 medical journals and ~50,000 clinical guidelines; designed for doctors and clinical researchers.
  • Study Snapshot + Ask Paper - automatic snapshot of any article (methodology, sample, results) and direct chat with the full-text PDF: AI highlights specific sections where information is found.
  • Results Timeline - chronological visualization of publications: each article is a point on a scale, and the size of the point is proportional to the number of citations, showing the historical development of the field.

Pros and Cons

Pros
  • Zero source hallucinations
  • Unique Consensus Meter
  • Database of 220M+ peer-reviewed articles
  • Filters by journal quartile and study type
  • Scholar Agent on GPT-5
  • 170+ universities offer Pro for free
Cons
  • No built-in editor or reference manager
  • Weak coverage of humanities
  • Meter does not weight study quality
  • No automatic filtering of retracted articles
  • Interface optimized for English only

Plans and Pricing

Free
$0/month
  • Unlimited basic search
  • 15 Pro messages per month
  • 3 Deep Searches per month
  • 10 Study Snapshots per month
  • Export to Zotero, Mendeley, EndNote
Deep
$45/month (billed annually)
  • All from Pro
  • 200 Deep Searches per month
  • For systematic reviews
Teams
$9.99/seat/month
  • Centralized billing
  • Individual accounts
  • Collaborative group work
Enterprise
Contact us
  • 200+ users
  • API access and LibKey integration
  • 170+ libraries already connected
Tip: Students get a 40% discount on paid plans; if your university is among the 170+ partners, Pro access may be completely free through an institutional license.

Comparison with Alternatives

Elicit is geared for structured systematic reviews with extracted data tables for 50-500 articles. Consensus excels where a quick visual answer to a binary question is needed - Elicit is stronger for formal construction of evidence tables.

Scite specializes in citation context: it shows how a specific article is cited - supportively, contradictorily, or neutrally. Consensus surpasses it in synthesizing responses from the entire database, while Scite is indispensable for verifying specific claims.

SciSpace (formerly Typeset) focuses on annotating PDFs and explaining complex concepts by paragraphs. Consensus leads in search and synthesis across the entire 220M+ article database; SciSpace is optimal for in-depth understanding of a single specific document.

Use Cases for Consensus

Literature Review
A graduate student receives a Deep Search report with a literature overview, gaps, and field chronology in 5 minutes instead of days of manual searching.
Clinical Decisions
A doctor checks the evidence base for a particular drug in 2 minutes through Medical Mode and gets a Consensus Meter with links.
Fact-Checking
A science journalist verifies press release statements against peer-reviewed literature without reading hundreds of articles manually.
Evidence-Based Products
Health-tech teams verify the scientific basis of product claims and quickly find cites for marketing.

Who Consensus is Suited For

  • Students and Graduates - accelerate literature reviews for theses and dissertations while maintaining academic rigor.
  • Researchers and Postdocs - quick scoping of new fields, hypothesis testing, and finding key works before in-depth reading.
  • Doctors and Clinicians - Medical Mode provides evidence-based answers from the top 1,000 medical journals in 2 minutes.
  • Science Journalists and Fact-Checkers - verify factual claims based on peer-reviewed literature, not web sources.
  • Policy and Business Analysts - substantiate decisions with scientific data without deep academic expertise.

How to Get Started with Consensus

  1. 1
    Sign Up - Go to consensus.app and create a free account; no card needed.
  2. 2
    Check Your University - Enter your corporate email: if your institution is among the 170+ partners, Pro access is activated automatically and free.
  3. 3
    Ask Your First Question - Formulate a binary question ("Does X cause Y?") in English and review the Consensus Meter with sources.
  4. 4
    Run a Deep Search - For a complex topic, use one of the 3 free Deep Searches a month: receive a structured report with a timeline and bibliography.
  5. 5
    Export Citations - With a single click, send links to Zotero, Mendeley, or EndNote; available at all pricing plans, including free.

Integrations

ZoteroMendeleyEndNoteRefWorksLibKeyChatGPT (ConsensusGPT)OpenAI APIMCP Server

Frequently Asked Questions

Is Consensus Free?

Yes, the free plan includes unlimited basic search, 15 Pro messages, 3 Deep Searches, and 10 Study Snapshots per month - no card required. Students can also access a 40% discount on paid plans.

How is Consensus Different from Perplexity AI?

Perplexity searches the entire web and quickly answers common questions. Consensus is restricted to peer-reviewed scientific literature - making it more reliable for scientific and medical inquiries, but not suitable for general topics.

How Does the Consensus Meter Work?

For a binary question, the system analyzes relevant peer-reviewed articles and displays a scale showing how many studies support the answer "Yes," "No," or "Maybe." It is important to note that the meter counts articles but does not weight their methodological quality.

Does Consensus Have an API?

Yes, API access is available at the Enterprise tier. In 2025, an MCP server (Model Context Protocol) was also released, allowing scientific database search integration into custom AI pipelines.

For researchers needing reliable answers with real sources, Consensus addresses the main pain point of academic search - the hours spent manually filtering through articles. Try the free plan or check if your university offers Pro access automatically.

← Back to "AI Search"