Consensus is an AI research workspace for finding, inspecting, and synthesizing scholarly literature.
Editorial score
Scholarly search combines semantic and keyword retrieval with detailed filters and citation-graph traversal.
Consensus does not access all research, and full-text availability depends on the source and access route.
Consensus is an AI research workspace focused on scholarly literature rather than general web search. It combines discovery, citation-backed synthesis, structured extraction, library workflows, and exports for researchers and teams. Its strongest use case is speeding up evidence discovery and review while keeping source inspection central; it is not a substitute for verifying original papers or conducting a complete systematic review.
Consensus is an AI workspace for scientific research. It retrieves scholarly literature and then analyzes and synthesizes the papers it finds. The product is designed for literature review work, including search, screening, extraction, organization, synthesis, and analysis of uploaded materials.
Consensus supports natural-language questions, keywords, Boolean operators, titles, authors, and DOIs through semantic and keyword retrieval. Filters cover publication date, study type, sample size, citation count, journal, field, country, open-access status, study design, publisher, and human or animal setting. Research Agent can plan multi-step tasks using search, DOI lookup, and citation-graph traversal.
Coverage is scholarly rather than general web search. Full text depends on the source and access route. Consensus does not access all research, and its Consensus Meter draws on only 5 to 20 top results for yes-or-no questions.
Generated claims include clickable citations that reveal supporting quotations from full text or abstracts. Citation grounding maps citations to exact paper quotes, section headers, and PDF highlights when available. In deep reviews, a checkmark indicates full-text use, while no checkmark indicates abstract-only use.
Retracted papers are flagged and excluded from AI-generated analyses. Citation inspection remains important because AI can misread a real source or summarize it incorrectly.
Consensus combines search, screening, extraction, synthesis, organization, collaboration, and export. Deep Research can create and run a search strategy and screen up to 1,000 papers. Table View structures fields such as population, methods, results, outcomes, sample size, duration, and country.
Library and Collections support saved papers, uploads, sharing roles, collection-specific queries, and chat across saved or uploaded materials. Outputs include PDF reports and CSV, RIS, and LaTeX exports. Extraction can be incomplete, and the available features do not provide complete audit-log or PRISMA-grade reproducibility controls.
Consensus suits researchers, students, educators, literature review users, clinicians, research groups, academic departments, developers, and internal AI teams. Library Collections and uploads also suit people working from a personal or shared corpus. The API and MCP provide ways to embed scholarly search into applications, agents, and copilot workflows.
It is not a general web search engine or a replacement for a complete systematic review. Medical and other high-stakes users should verify findings against the original literature.
Where the product fits — and where to be careful
Strong fit for accelerating discovery, screening, extraction, and synthesis while retaining researcher verification.
Useful for evidence discovery and review when paired with professional judgment and primary-source verification.
Clickable citations expose supporting quotations, while Citation Grounding can link claims to exact quotes, section headers, and PDF highlights. Full-text versus abstract-only use is distinguished in Deep reviews, though source verification remains necessary.
Consensus supports search strategy creation, screening, structured extraction, synthesis, Collections, sharing, uploads, and CSV, RIS, LaTeX, and PDF outputs. It lacks established evidence of complete reproducibility controls for systematic reviews.
The platform offers scholarly semantic and keyword search, extensive filters, citation-graph traversal, medical coverage, and a corpus reported at more than 400M scholarly sources. Coverage and full-text access remain conditional, and not all research is included.
User-review evidence was insufficient to score usability, support, or value.
Consensus has Free, Pro, Deep, Teams, and Enterprise options. Usage limits vary by plan.
| Plan | Price | Includes |
|---|---|---|
Free | $0 monthly | Includes unlimited Papers searches, limited Pro messages, 3 Deep reviews per month, Study Snapshots, and included API/MCP calls. |
Pro | $20/month or $144/year monthly or annual | Includes unlimited Pro messages and 15 Deep reviews per month. |
Deep | $65/month or $540/year monthly or annual | Includes 200 Deep reviews per month. |
Teams | $20 per seat/month or $240 per seat/year monthly or annual | Available for up to 200 seats. |
Documented connections and access methods for Consensus.
Recent product and company updates
Microsoft 365 Copilot availability
Consensus became an official Microsoft Marketplace app for Microsoft 365 Copilot.
SourceShowing 10 of 22 sources · 12 additional sources analyzed
Rate Consensus