Overview

Consensus is an AI-powered academic search engine that enables users to ask natural language research questions and obtain evidence-based summaries drawn from over 200 million peer-reviewed scientific documents. Unlike general AI tools, each answer is grounded in cited research and traceable to the original paper. Consensus offers tools like the Consensus Meter to gauge agreement across studies, Pro Analysis (formerly Copilot) for flexible synthesis, and advanced filters for precise literature exploration.

Features and Capabilities

• **Corpus and Coverage: **Consensus searches a corpus of over 200 million documents including peer‑reviewed journals, conference papers, and preprints, aggregated from multiple academic sources and its own web crawl.
• **AI‑powered Synthesis: **Uses hybrid search (semantic embeddings + keyword BM25) to retrieve relevant papers, then generates cohesive summaries tied directly to citations.
• **Consensus Meter: **For binary “yes/no” queries, analyzes the first 20 relevant papers and classifies consensus results as “yes,” “no,” or “possibly,” with transparent sources and extracted quotes.
• **Pro Analysis (formerly Copilot): **Enables command‑based synthesis like drafting outlines or answering complex questions, while maintaining citation integrity.
• **Study Snapshot & Reference Export: **Displays population, methods, sample size, and more for quick evaluation; supports exporting references to CSV or RIS for EndNote, Mendeley, and Zotero.
• **Advanced Filtering: **Filter results by study type (RCT, meta‑analysis), population (human, animal), journal prestige, sample size, and other metadata for tailored exploration.