Consensus MCP

@consensus what does recent research say about the effectiveness of intermittent fasting for weight loss compared to continuous calorie restriction?

Based on a search of peer-reviewed literature, Consensus found 8 relevant studies published between 2019-2024. The meta-analysis by Patikorn et al. (2021) showed intermittent fasting and continuous calorie restriction produced similar weight loss outcomes (mean difference: -0.4 kg, not statistically significant). However, Harris et al. (2023) noted intermittent fasting had higher dropout rates in trials lasting >12 weeks. The consensus meter shows 'possibly effective' with moderate evidence quality. Three studies reported improved insulin sensitivity markers independent of total weight lost.

This example shows Consensus's core strength: synthesizing academic research with citation-backed summaries. The MCP searches a database of scientific papers, not general web content. Results depend on how the query maps to available literature—niche topics may return fewer studies. OAuth is required to access the full paper database and consensus metrics.

@consensus find studies on the relationship between screen time and adolescent mental health, then summarize the main findings and any conflicting results

Consensus retrieved 12 studies examining screen time's impact on adolescent depression and anxiety. The synthesis shows mixed evidence: Twenge & Campbell (2018) found a positive correlation between social media use and depressive symptoms (r=0.15), while Orben & Przybylski (2019) reported the effect size was negligible when controlling for other variables (β=0.04). A 2022 longitudinal study by Valkenburg noted the relationship was bidirectional—teens with existing anxiety increased screen time, which then modestly worsened symptoms. The consensus meter indicates 'uncertain' due to methodological heterogeneity across studies.

This demonstrates Consensus's ability to surface contradictory findings rather than forcing a single narrative. The AI synthesizes the MCP's results into a coherent summary, highlighting where the research community disagrees. Useful for due diligence before making claims based on 'the science says' assumptions. Note that paper access may be limited by your OAuth scope.

@consensus I'm writing a grant proposal about urban green spaces and public health outcomes—what's the strongest evidence I can cite for cardiovascular benefits?

Consensus identified 6 high-quality studies linking urban green space exposure to cardiovascular health. The strongest evidence comes from Twohig-Bennett & Jones (2018), a systematic review of 290,000+ participants showing green space access correlated with reduced diastolic blood pressure (-1.54 mmHg) and heart rate (-2.57 bpm). Gascon et al. (2016) found a 16% reduction in cardiovascular mortality for those living within 300m of green spaces. The consensus meter rates this 'likely effective' with high evidence quality. However, most studies are observational—only 2 RCTs exist, both with small sample sizes (<200 participants).

This example shows how Consensus supports real-world research tasks by filtering for evidence quality and study design. The MCP returns citation details suitable for grant writing or literature reviews. The AI's synthesis highlights both the strength of the evidence and its limitations (observational vs. experimental). Be aware that Consensus focuses on published papers—preprints and grey literature may not be included.

When OAuth-gated research tools fit quarterly planning cycles

A 6-person product team runs quarterly roadmap reviews and needs to pull academic or industry research to validate feature bets. Consensus MCP uses OAuth2, so each team member authenticates once and the workspace inherits their search scope. This works if your planning cycle is slow enough to absorb the OAuth setup overhead—figure 20 minutes per person the first time, then zero friction after that. The trade-off: if you're running daily scrums where research questions pop up ad-hoc, the auth boundary adds latency. If your roadmap reviews happen every 12 weeks and involve the same 4-6 people, the OAuth model pays off. You get scoped access without API key sprawl, and the MCP handles token refresh in the background. Best fit is teams that batch research questions into scheduled sessions rather than answering them in real time.

Why OAuth matters for client-facing research lookups

A 3-person customer success team fields questions from enterprise clients who want evidence that a feature request aligns with industry best practice. Consensus MCP's OAuth2 flow means each CSM authenticates with their own credentials, and the workspace logs which team member ran which search. This audit trail matters if your clients ask how you sourced a claim or if compliance wants to know who accessed what. The downside: if your CS team rotates frequently or you onboard contractors for overflow, you're re-authenticating people every few weeks. The threshold is team stability. If your CS roster is stable for 6+ months and you need per-user attribution, OAuth is the right call. If you're staffing up and down or using shared logins, the auth overhead outweighs the benefit. Works best for small, stable teams handling high-stakes client questions.

When zero-tool MCPs still add value through scoped auth

A 2-person investment analyst team evaluates early-stage startups and needs to verify technical claims in pitch decks. Consensus MCP shows zero tools in the directory, which usually means the vendor hasn't exposed granular functions yet—but the OAuth2 scope still gates access to a research corpus. This setup works if the MCP is really a thin wrapper around a search API and your team already knows the query syntax. You're trading tool discoverability for faster auth. The risk: if the MCP doesn't expose tools, your AI agents can't auto-discover what's possible, so you're writing raw queries by hand. If your analysts are comfortable with that and the corpus is narrow enough to memorize, the OAuth boundary is worth it. If you need the AI to explore capabilities on its own, wait until the vendor ships a tool manifest. Best for teams that already know exactly what they're searching for.