Consensus

Consensus is a search engine that answers research questions using findings from peer-reviewed papers rather than the open web. You type a question in plain language — “does intermittent fasting improve insulin sensitivity?” — and it returns relevant studies alongside a one-line summary of what each one found. It draws on the Semantic Scholar database, which indexes on the order of 200 million papers across disciplines.

What sets it apart from Google Scholar or PubMed is synthesis. Instead of handing you a list of titles to open one by one, Consensus extracts the key finding from each study and, for yes/no questions, aggregates them into a Consensus Meter showing how many studies support, oppose, or are unsure about a claim. It sits closer to tools like Elicit and Scite, though it leans more toward quick evidence checks than full systematic-review workflows.

The company is a startup founded in 2021 by Eric Olson and Christian Salem. Underlying summaries are generated with large language models layered on top of the retrieved abstracts.

Key Features

• Natural-language questions answered from peer-reviewed literature
• Consensus Meter aggregating agree/disagree across studies for yes/no questions
• Study Snapshot showing population, sample size, and methodology at a glance
• Filters for study design (RCT, meta-analysis), publication year, and journal
• GPT-based “Copilot” summaries synthesizing findings across several papers
• Direct links and citations for every study surfaced

Pricing

• Free: Unlimited searches, but a limited monthly allowance of AI Copilot and Consensus Meter credits
• Premium: Around $9 per month (billed annually) for unlimited AI features and GPT-4-class synthesis
• Enterprise / Teams: Custom pricing with shared workspaces and admin controls

Best For

Researchers, clinicians, grad students, and evidence-minded writers who need to know what the literature actually says on a specific question — and want citations, not a chatbot’s unsourced guess.

Limitations

Coverage is strongest in medicine and the social sciences and thinner in fields like pure math or engineering. The Consensus Meter is a helpful signal, not a verdict — it reflects abstract-level summaries and cannot weigh study quality the way a trained reviewer would. For anything high-stakes, treat it as a starting point and read the source papers.