Home · ← Semantic Scholar | 5/8 | Elicit →

Consensus

AI-summarized evidence — read with the abstract still open

Overview

Consensus

A synthesis layer over the literature — useful, and itself a claim you have to check.

What it is

What it is

consensus.app — synthesizes findings across many papers into a quick evidence snapshot for a question you ask in plain language.

Where it fits

Where it fits

D1·1Environment check
D1·2AGENTS.md
D1·3Literature tools
D1·4Referee skill
D1·5Subagents
D2·1Hypothesis
D2·2Design
D2·3Data lab
D2·4Results
D2·5Manuscript
D2·6Governance

Day 1's literature exercise — plus D1·4/D1·5, since literature-scout accepts a Consensus result the same way it accepts a Semantic Scholar one, and D2·5, since a citation backed this way can end up in the manuscript.

Access & cost

Access & cost

consensus.app, sign up with email. Free tier covers the course; Premium is ≈ €10–15/mo if you need more beyond it.

In the course

D1·3 exercise

Ask Consensus the same underlying question as your Semantic Scholar results (e.g. "does AI exposure reduce hours worked?") and compare what comes back to the actual paper's abstract.

In the course

The check this teaches

If the synthesis doesn't match what the abstract itself says, trust the primary source — the same discipline as literature-scout's “unverified” rule, applied to a synthesis tool instead of a raw citation.

Limits

Where it's weaker than Semantic Scholar

It's a synthesis layer, not a citation graph — good for “what does the literature broadly say,” not for finding or verifying a specific paper exists.

Limits

Where it's weaker than Scite

Doesn't track whether a specific paper's findings have been supported or contradicted by later citing work — that's a different, narrower question Scite answers.

In the course

Used for positioning, too

The Day-1 afternoon block (“literature & positioning”) leans on Consensus (and Elicit) to find the gap in the literature for your own research question, not just the shared case.

In the course

Feeds the manuscript, too

literature-scout accepts a Consensus result you've pasted into the chat as valid support for a citation — the same standing as a Semantic Scholar search result. Anything Consensus surfaces can end up cited in the manuscript (D2·5), and gets cross-checked again by the referee skill's citation check (D1·4).

Alternative

The build-vs-buy alternative

Elsevier's LeapSpace bundles this kind of synthesis with full-text access across several publishers under one institutional login — worth a quick check with the library before buying separate seats for Consensus/Elicit/Scite.

Takeaway

What stays constant either way

Whichever synthesis tool you use, the rule doesn't change: an AI summary of a paper is an unverified claim until you've opened the paper.

Recap

Recap

Back to D1·3's decision graph on the home page.