AI for Research & Knowledge

Consensus — The Complete AI Research Guide

Consensus is an AI search engine for research questions. Ask 'Does exercise reduce depression?' and Consensus searches 200 million academic papers, synthesises the findings, and tells you what the research actually says — with citations. Free tier available. Premium: $8.99/month.

AI Evidence Search200M+ papersConsensus MeterFree tier availablePremium: $8.99/monthLast reviewed: April 2026

What is Consensus?

Consensus is an AI-powered search engine that answers research questions using peer-reviewed academic literature. Unlike a general search engine that returns web pages, Consensus searches 200 million academic papers and synthesises their findings into a direct answer with citations. You ask a question the way you would ask a researcher, and Consensus tells you what the evidence says.

The signature feature is the Consensus Meter — a visual indicator showing whether the academic literature broadly supports, contradicts, or is divided on a claim. It distinguishes between: "the evidence strongly supports this", "the evidence is mixed", and "the evidence contradicts this." This saves hours of reading through conflicting study summaries trying to assess the weight of evidence.

Consensus is the fastest route from a research question to a reliable evidence-based answer. It is not designed for the deep extraction work of systematic reviews (use Elicit for that), but for quickly checking what the research says on a specific question — before a decision, before citing a claim, before writing a recommendation.

Who Consensus is for

Researchers who need quick evidence checks across many questions. Students and academics checking claims before writing. Healthcare professionals wanting a rapid synthesis of the evidence on a treatment question. Journalists fact-checking health and science claims. Policy analysts assessing the evidence base for an intervention. Anyone who needs to know what academic research says on a specific question without reading dozens of papers.

Getting started

Go to consensus.app. No sign-up required for basic searches. For the full experience including the Consensus Meter and GPT-4 synthesis, create a free account. Enter your research question as a complete question — "Does intermittent fasting improve metabolic health in adults?" — not keywords. Consensus returns a synthesised answer plus the top supporting papers with links.

12 Consensus use cases

Evidence check before citing
Before citing a claim in your writing: search it as a question on Consensus. Check the Consensus Meter. If it shows 'strongly supported', cite confidently with the papers Consensus found. If it shows 'mixed' or 'contradicted', investigate the nuance before making a strong claim.
Quick literature survey
For a new topic you are entering: run 5–10 key questions on Consensus before doing deeper research. This gives you the lay of the land — which questions have strong consensus, which are contested, which are understudied — in 15 minutes rather than days.
Check an intervention
'Does [intervention] work for [outcome] in [population]?' Consensus tells you: what the overall body of evidence says, how large the effects are, and how consistent findings are across studies. Use this to decide whether to investigate further.
Compare two approaches
Run both 'Does approach A improve [outcome]?' and 'Does approach B improve [outcome]?' Compare the Consensus Meters and the strength of evidence for each. This is useful for evidence-based decision-making between two options.
Fact-check a health claim
A patient, client or colleague asserts [health claim]. Search it on Consensus before responding. If the evidence does not support it, you have citations to point to. If it does, you can confirm it with confidence. Faster and more reliable than a Google search.
Build an evidence base for a recommendation
You need to recommend [policy / intervention / practice change]. Run 10 specific questions on Consensus covering: effectiveness, safety, cost-effectiveness, implementation factors. Compile the results into a brief evidence summary. Each claim has citations from peer-reviewed literature.
Pre-research scoping
Before beginning a research project: search your core hypothesis on Consensus. If there is already strong consensus on your question, your contribution needs to be methodological novelty or an underserved population. If evidence is mixed or absent, that gap is your opportunity.
Student essay preparation
Before writing an essay or thesis chapter: search your key claim or argument on Consensus. Collect the supporting papers into a reading list. Know the state of the evidence before you start writing, rather than discovering it as you go.
Deep Research mode
On paid plans, use Consensus Deep Research for comprehensive investigations. Enter a complex question. Consensus searches more broadly, synthesises across more papers, and produces a detailed report with structured sections. Useful for complex policy or medical questions requiring multi-faceted answers.
Export to reference manager
After a Consensus search, export the supporting papers to your reference manager. All major reference managers are supported. This turns a 2-minute evidence check into a citable bibliography entry that can go straight into your writing.
Check claims in a document
Working through a report or paper with multiple factual claims: search each key claim on Consensus. Flag any where evidence is mixed or contradicted. This is a quick way to quality-check AI-generated or student-written content for factual reliability.
Newsletter or content research
Before publishing a piece on a health, science or social topic: verify your key claims on Consensus. This takes 10 minutes and means you are not inadvertently spreading contested or contradicted claims. Cite the evidence for key points.

Tips

Frame questions precisely. "Does exercise help depression?" returns broader results than "Does aerobic exercise reduce depressive symptoms in adults with major depressive disorder?" The more specific your question, the more specific and useful the Consensus Meter result.

Read the caveats on the Meter. Consensus adds important context — "studies are mostly small", "evidence is mostly from high-income countries", "most studies are short-term". These qualifications matter as much as the headline finding.

Technical background

Consensus was founded in 2021 and searches over 200 million academic papers, per the official Consensus website. The Consensus Meter uses AI classification to categorise paper findings as supporting, contrasting or inconclusive, and aggregates across results to produce the summary indicator. GPT-4 synthesis (available on paid plans) produces the written summary above the paper results.

Pricing (verified April 2026)

  • Free: 10 research analyses/month, basic features
  • Premium: $8.99/month (annual) or $10/month — unlimited searches, GPT-4 synthesis, Deep Research, export
  • Enterprise: Custom — team access, API, institutional licensing
Primary sources