Consensus Review: Fast Research Triage, Not a Systematic Review
Consensus is an AI search tool that surfaces and summarizes findings from 250M+ peer-reviewed papers, using a Consensus Meter and Snapshot to show agreement levels and per-position quality signals like methods and journal SJR. It's useful for quick, cited answers to yes/no research questions and literature screening, but coverage gaps, abstract-only records for non-partner publishers, and its triage-level quality signals mean it doesn't replace a formal systematic review or careful reading of source studies.
What it is
Consensus is an AI search engine that queries a corpus of 250M+ peer-reviewed research papers to answer research questions, surface supporting/opposing studies, and summarize findings with citations. It licenses full-text content from publishers including Wiley, AAAS, Taylor & Francis, Sage, ACS, and APA where available, falling back to metadata/abstracts otherwise (per consensus.app/home/features/full-text/). Note: the homepage previously cited 220M+ papers; the current official figure is 250M+.
How it works
Ask a yes/no research question and the Consensus Meter renders a visual agree/disagree read across relevant papers. Consensus Meter 2.0 (launched Feb 6, 2025) improved on this with a per-position Consensus Snapshot showing four quality indicators - Recency (average publish date), Methods (count of meta-analyses, systematic reviews, RCTs), Journals (average SJR score), and Citations (total citation count) - plus a Mixed category for nuanced findings and badges for which position wins a quality category. This addresses a real flaw in the original Meter, where a single case report could count the same as a Cochrane systematic review. As of July 8, 2026, advanced search adds 19 study-design filters, 44 publisher filters, and a human/animal setting filter. Users can also chat with or analyze up to 5 selected/uploaded papers (added Oct 2025) and use a Table View to compare population, methods, results, sample size, and duration across studies. Even with these upgrades, the Meter/Snapshot remains a triage signal, not a substitute for reading methods and results yourself - SJR scores and citation counts are proxies for influence, not guarantees of validity.
Pricing
Per the live pricing page (consensus.app/pricing, checked 2026-07-22): Free is $0/mo with basic paper search, 15 Pro messages/month, and up to 3 Deep reviews/month. Pro is $12/mo or $144/year (saving $96/yr), adding unlimited Pro messages, 15 Deep reviews/month, and unlimited access to all research tools (search, analysis, citation graph, and more). Deep is $45/mo or $540/year (saving $240/yr), with unlimited Pro messages and 200 Deep reviews/month. Students and faculty with a valid school email, and US healthcare professionals with a valid NPI number, can get up to 40% off. Note: some third-party or help-center pages still list older pricing (e.g., Pro at $15/mo, Deep at $65/mo) and older feature names - the live pricing page above is the authoritative, current source, and both prices and terminology have since changed.
Real limitations
Coverage is broad but not exhaustive or reproducible, and Consensus is not a substitute for a formal PRISMA-style systematic review. Full text is only available for open-access or partner/licensed publishers; other records rely on metadata or abstracts alone, which limits synthesis depth. The tool is comparatively weak for arts and humanities topics and anything not well indexed in mainstream peer-reviewed databases. Quality indicators in the Consensus Snapshot help contextualize findings but still require human judgment - they do not replace actually reading and appraising the underlying studies, especially for clinical or other high-stakes decisions.
Who it's for
Consensus fits researchers, students, and clinicians who want a fast, cited first pass on a yes/no research question or a way to screen literature with some quality context attached. It suits people who want AI-assisted synthesis rather than a citation manager or exhaustive review tool. It's a poor fit for anyone needing the completeness and reproducibility of a true systematic review, or working primarily in humanities fields where its coverage is thinner.
Sources
- Consensus Meter 2.0's Snapshot adds real quality context (recency, methods, journal SJR, citations) beyond a simple yes/no split
- Draws on a large, continuously expanding corpus (250M+ papers) with licensed full text from major publishers where available
- Free tier and tiered Pro/Deep plans with clear per-feature limits, plus up to 40% discounts for students, faculty, and US healthcare professionals
- Advanced filters (19 study designs, 44 publishers, human/animal setting) and multi-paper chat/comparison tools support more targeted screening
- Not exhaustive or reproducible enough to replace a formal systematic review process
- Full text is limited to open-access/partner publishers; many records rely on abstracts only
- Consensus Meter and Snapshot are triage aids, not validity guarantees - SJR and citation counts are proxies
- Weaker coverage for arts/humanities and non-indexed proprietary material
- Third-party sources still list outdated pricing/names, creating potential confusion despite the live page being authoritative