---
canonical_url: "https://docs.composio.dev/toolkits/consensus_mcp"
markdown_url: "https://docs.composio.dev/toolkits/consensus_mcp.md"
product: "platform"
version: null
last_updated: null
---

# Consensus MCP

Consensus MCP lets agents search peer-reviewed academic research and return papers, citations, study metadata, and plan-dependent analysis fields.

- **Category:** model context protocol
- **Auth:** DCR_OAUTH
- **Composio-managed OAuth available?** No
- **Tools:** 1
- **Triggers:** 0
- **Slug:** `CONSENSUS_MCP`
- **Version:** 20260910_00

## Tools

### Search

**Slug:** `CONSENSUS_MCP_SEARCH`

Search over 220 million peer-reviewed academic papers.

Use this tool when the user asks about research, studies, evidence,
scientific findings, clinical trials, medical research, or academic
literature. Covers Semantic Scholar, PubMed, Scopus, and ArXiv.

Returns paper titles, authors, abstracts, citation counts, journal
quality scores, and direct URLs to each paper.
All filters are optional — only apply them when the user explicitly asks.

Example query → filter mappings:
- "what does the evidence say about metformin for longevity?"
    → query="metformin longevity mortality",
      medical_mode=True, exclude_preprints=True
- "recent peer-reviewed RCTs on SGLT2 inhibitors for heart failure"
    → query="SGLT2 inhibitors heart failure outcomes",
      year_min=2022, exclude_preprints=True, study_types=["rct"]
- "long-term studies of statin side effects (1+ year)"
    → query="statin adverse effects long-term",
      duration_min=365, medical_mode=True
- "what is dark matter?"
    → query="dark matter"  (no filters — general question)
- "coffee and cognition, excluding top-tier Q1 journals"
    → query="coffee cognition", sjr_min=2

#### Input Parameters

| Parameter | Type | Required | Description |
|-----------|------|----------|-------------|
| `page` | string | No | Zero-indexed result page. Set ONLY when the user explicitly asks for more or next results. Pages above 0 are available with paid plans (Pro, Deep, Teams, Enterprise) or an Enterprise API key. |
| `human` | string | No | Restrict to human studies only. Set ONLY when the user explicitly asks for human / clinical / patient studies (or to exclude animal / in-vitro work). Do NOT set by default for medical, health, or clinical questions — the corpus already includes the relevant human evidence and filtering here may exclude valuable translational, mechanism, or preclinical context. |
| `query` | string | Yes | Search query. Be specific and use academic terminology. |
| `domain` | string | No | Comma-separated academic field short codes, not email or web domains. Set ONLY when the user explicitly asks to limit results to specific academic disciplines or fields. Valid codes: med, bio, cs, chem, psych, phys, mat, eng, env, bus, econ, math, poli, agri, edu, soc, geol, geog, hist, art, philo, law, ling. Examples: 'medical papers' -> 'med', 'biology and computer science' -> 'bio,cs'. Do NOT set for general research questions. |
| `country` | string | No | Comma-separated ISO 3166-1 alpha-2 country codes (e.g. 'us', 'gb') to limit to those countries of study. Set ONLY when the user explicitly names a country or region of study. Do NOT infer from the query's language or topic. Unknown codes are ignored. |
| `sjr_max` | string | No | Maximum journal ranking quartile (1=Q1 top → 4=Q4). Set ONLY when the user explicitly asks for top-tier / Q1 / high-impact-journal results. Do NOT set by default — many landmark and niche-field studies publish outside Q1. |
| `sjr_min` | string | No | Minimum journal-ranking quartile floor (1=Q1 top -> 4=Q4). Keeps journals at or below this quartile rank, i.e. it EXCLUDES the very top tiers. Set ONLY when the user explicitly asks to exclude top-tier / Q1 journals. For 'top-tier / Q1 / high-impact', use sjr_max instead, never sjr_min. Rarely set. |
| `year_max` | string | No | Maximum publication year. Set ONLY when the user asks for historical research before a specific date. Rarely set. |
| `year_min` | string | No | Minimum publication year. Set ONLY when the user asks for recent / latest / 'in the last N years' research. Do NOT set by default. |
| `month_max` | string | No | Maximum publication month, paired with year_max. Set ONLY when the user specifies an upper-bound month or exact date such as 'through June 2022'. Do NOT set without year_max. |
| `month_min` | string | No | Minimum publication month, paired with year_min. Set ONLY when the user specifies a lower-bound month or exact date such as 'after March 2020'. Do NOT set without year_min. |
| `page_size` | string | No | Number of results to return per page. Set ONLY when the user explicitly asks for a specific page size. If omitted, the caller's tier default is used. |
| `controlled` | string | No | Restrict to controlled studies. Set ONLY when the user explicitly asks for controlled studies or controlled trials. Do NOT set by default. |
| `open_access` | string | No | Restrict to open-access papers. Set ONLY when the user explicitly asks for open-access or free-to-read papers. Do NOT set by default. Note: 'open source' is NOT 'open access'. |
| `study_types` | string | No | Filter by study design. Set ONLY when the user names a design ('RCT', 'meta-analysis', 'systematic review') or asks for 'best' / 'highest-quality' / 'rigorous' evidence (then use ['rct', 'meta-analysis', 'systematic review']). Do NOT set for general research questions — narrowing by design excludes valid evidence. |
| `citation_min` | string | No | Minimum citation count. Set ONLY when the user explicitly asks for highly-cited or influential work. Do NOT infer from words like 'seminal' or 'foundational'. Do NOT set by default. |
| `duration_max` | string | No | Maximum study duration in DAYS. Rarely set. |
| `duration_min` | string | No | Minimum study duration in DAYS. Set when the user asks for longitudinal / long-term / multi-year studies. Typical thresholds: 30 (month), 365 (year), 730 (2+ years). |
| `journal_name` | string | No | Preferred journal name (e.g. 'Nature'). Results from matching journals are surfaced higher; this does not exclude other journals. Set ONLY when the user explicitly asks to prefer a specific journal. Do NOT set by default. |
| `medical_mode` | string | No | Filter to top medical journals and clinical guidelines (~8M documents). Set to true for clinical / medical / evidence-based-medicine questions (e.g. drug efficacy, treatment comparisons, diagnostic criteria, NEJM/JAMA/Lancet-tier evidence requests). Pairs well with exclude_preprints=true. |
| `publisher_name` | string | No | Comma-separated publisher display names to filter by. Set ONLY when the user explicitly asks for papers from a specific publisher (e.g. 'Elsevier', 'Nature', 'PLOS ONE'). Do NOT set by default. |
| `sample_size_min` | string | No | Minimum participant count. Set ONLY when the user asks for large or well-powered studies. Do NOT set by default — many high-quality studies are small by design. |
| `exclude_preprints` | string | No | True to exclude preprints and return peer-reviewed papers only. Set when the user asks for 'peer-reviewed', 'published', 'vetted' research or explicitly 'no preprints'. |
| `include_full_text_chunks` | string | No | Return query-relevant full-text excerpts. Set ONLY when the user explicitly asks for excerpts from full text and has a paid plan (Pro, Deep, Teams, Enterprise). |