A citation count tells you a paper was talked about — not whether it was supported, contradicted or merely mentioned. And a literature's 'consensus' on a question is usually something you assemble painstakingly by hand. Scite and Consensus each attack one of these problems with AI, and they're worth knowing precisely because they answer different questions.
Scite: how was this paper cited?
Scite's 'smart citations' classify the context of each citation to a paper: does the citing sentence provide supporting evidence, contrasting evidence, or a mere mention? Before you build an argument on a key study, its Scite report tells you whether later work replicated it, disputed it or ignored it — the difference between citing a load-bearing finding and citing a retracted-in-spirit one. It also flags formal retractions and editorial concerns, which alone justifies a pre-submission pass over your reference list.
Consensus: what does the literature say about X?
Consensus is an AI-powered search over academic papers that answers question-shaped queries — 'does remote work improve productivity?' — with a synthesis of findings and, for suitable yes/no questions, a consensus meter summarising the balance of evidence. It's built for the orientation stage: within minutes you know whether a question is settled, contested or barely studied — which is exactly the input a research gap hunt needs.
When to reach for which
- Vetting a cornerstone reference before your argument depends on it → Scite.
- Checking your reference list for retractions/disputes pre-submission → Scite.
- First orientation on a research question you might pursue → Consensus.
- A quick read on where the evidence leans before designing a study → Consensus, then verify by reading the top papers.
Both tools depend on what their corpora cover and how well AI classified or summarised it — coverage gaps and misclassifications happen. Treat both as triage: they tell you where to look and how hard to look, never what to write. Claims in your thesis rest on papers you read, not on a meter.
Both slot into the verification stage of the AI research toolchain. For turning verified evidence into a publishable argument, our publication mentoring picks up where the tools stop.
Frequently asked
Is Scite's supporting/contrasting classification reliable?+
It's good triage but imperfect — citation contexts are classified by AI and nuance gets lost. Use it to prioritise which citing papers to read; confirm the characterisation in the actual text before repeating it.
Can I cite the Consensus meter in my thesis?+
No — cite the underlying studies. The meter is an orientation device over a subset of the literature, not a systematic synthesis; examiners will (rightly) treat it as neither.
Are Scite and Consensus free?+
Both are freemium: limited free access with subscriptions for full features. Some institutions license Scite — check your library before paying personally.
phdguide's mentors are senior academics, former supervisors, statisticians and publication specialists with 25+ years of combined experience guiding MBA, MPhil and PhD scholars from topic to viva.
Ethical, compliant guidance: We provide academic support, mentoring, analysis, editing and structuring — not authorship. Your work stays compliant with university policies.