AI-powered research platform that provides answers grounded in over 1.6 billion citation statements extracted from 280M+ peer-reviewed articles, preprints, books, patents, and datasets, using Smart Citations to classify each citation as supporting, contradicting, or mentioning.
AI-powered research platform that provides answers grounded in over 1.6 billion citation statements extracted from 280M+ peer-reviewed articles, preprints, books, patents, and datasets, using Smart Citations to classify each citation as supporting, contradicting, or mentioning.
scite is an AI-powered research platform that fundamentally changes how researchers, students, and professionals evaluate scientific literature. Instead of simply counting how many times a paper has been cited, scite analyzes over 1.6 billion citation statements extracted from more than 280 million peer-reviewed sources to classify each citation as supporting, contradicting, or mentioning the referenced claim. This Smart Citation technology, built on deep learning models trained on millions of annotated citation contexts, gives users an evidence-based view of how scientific findings hold up over time.
The platform's scite Assistant is a conversational AI that answers research questions using exclusively peer-reviewed literature, with every claim backed by inline citations linking directly to Smart Citation context. Unlike general-purpose AI chatbots, the Assistant is designed to minimize hallucination by grounding all responses in indexed academic sources. In internal evaluations, over 90% of Assistant-generated citations were verified as accurately reflecting the source material.
scite's Reference Check feature allows authors to upload manuscripts and receive automated audits flagging any cited papers that have been retracted, issued editorial notices, or contradicted by subsequent research. This pre-submission tool has been used to review over 500,000 manuscripts since its launch, helping researchers avoid citing unreliable work before peer review. Custom Dashboards let users build visual citation landscape overviews for any topic, substance, or research question, quickly identifying patterns of scientific consensus or controversy across hundreds or thousands of papers.
The platform's full-text search capability extends beyond abstracts and metadata through direct licensing agreements with over 30 publishers, including Wiley, SAGE, and Taylor & Francis, surfacing findings buried in methods sections, supplementary materials, and discussion paragraphs that abstract-only tools miss entirely. scite's browser extension integrates with PubMed, Google Scholar, and major journal websites, overlaying real-time Smart Citation badges that display supporting and contradicting citation counts without requiring users to leave their existing research workflow.
As of early 2026, scite serves over 750,000 registered users across more than 1,200 institutional subscribers worldwide, including universities, pharmaceutical companies, and government research agencies. The platform processes approximately 2 million search queries per month and has classified citation contexts across all major academic disciplines, from biomedical sciences and engineering to social sciences and humanities. With the recent launch of MCP (Model Context Protocol) integration and a developer API, scite is expanding beyond its web interface to power citation intelligence in third-party research tools, AI agents, and institutional knowledge systems.
Was this helpful?
scite's deep learning model analyzes each citation context and classifies it as supporting, contradicting, or mentioning, with confidence scores displayed for each classification. Built on training data of millions of expert-annotated citation contexts, the system processes citation statements from over 280 million sources, making it the largest classified citation database available to researchers.
Unlike tools limited to abstracts and metadata, scite searches the full text of articles through direct licensing agreements with Wiley, SAGE, and 30+ other publishers. This means researchers can find relevant findings mentioned in methods sections, supplementary materials, and discussion paragraphs that would be invisible to abstract-only search engines.
The conversational AI assistant generates answers exclusively from peer-reviewed literature, with every claim linked to its source paper and Smart Citation context. Users can click through to see the exact citation statement, verify whether it supports or contradicts the referenced finding, and follow links to the original paper for further reading.
Authors upload a manuscript and receive an automated audit of every reference, flagging papers that have been retracted, issued editorial notices, or contradicted by later work. This pre-submission tool has processed over 500,000 manuscripts, helping authors catch problematic citations before peer reviewers or editors identify them.
scite offers an API for institutions and developers to build custom tools on top of its citation intelligence, plus a recently introduced MCP (Model Context Protocol) integration that allows AI agents and automated research workflows to query scite's Smart Citation database programmatically, enabling citation-aware AI applications beyond scite's own interface.
$0/month
$20/month (billed monthly) or $144/year ($12/month billed annually)
Custom pricing based on institution size (typically $2,000–$15,000+/year)
Ready to get started with scite?
View Pricing Options →Researchers conducting systematic reviews can use scite's Smart Citations to quickly identify which key findings in their domain have been supported or contradicted across hundreds of subsequent studies, dramatically reducing the time spent manually tracing citation chains and improving the overall rigor of their evidence synthesis.
Before submitting a paper for peer review, authors can upload their manuscript to scite's Reference Check tool to automatically flag any cited papers that have been retracted, contradicted, or issued editorial notices since they were originally read, catching problematic references that could undermine the manuscript's credibility.
Grant writers can use the scite Assistant to quickly assemble cited, peer-reviewed evidence supporting their proposed hypotheses, ensuring that every claim in the proposal is backed by literature that has itself been validated by subsequent research through Smart Citation analysis.
Journalists covering scientific topics and policy analysts evaluating evidence-based claims can use scite to verify whether a study cited in press releases or policy documents has been broadly supported or contradicted by the wider research community, adding rigor to their reporting and recommendations.
Pharmaceutical companies and clinical research teams can use scite's Custom Dashboards to map the citation landscape around a drug compound, treatment approach, or biomarker, quickly identifying which foundational studies have been replicated and supported versus those that remain contested or contradicted.
Professors can use scite to teach students how to critically evaluate sources by demonstrating the difference between papers that have been widely supported versus those that have been contradicted, instilling evidence-based research habits that go beyond simple citation counting.
We believe in transparent reviews. Here's what scite doesn't handle well:
Weekly insights on the latest AI tools, features, and trends delivered to your inbox.
scite has launched an MCP (Model Context Protocol) integration, enabling developers to connect scite's Smart Citations database and search capabilities to AI agents and automated research workflows. The platform has also expanded its publisher partnerships beyond 30 providers, grown its indexed source base past 280 million documents, and introduced confidence scores on Smart Citation classifications to help users gauge reliability at a glance.
No reviews yet. Be the first to share your experience!
Get started with scite and see if it's the right fit for your needs.
Get Started →Take our 60-second quiz to get personalized tool recommendations
Find Your Perfect AI Stack →Explore 20 ready-to-deploy AI agent templates for sales, support, dev, research, and operations.
Browse Agent Templates →