Scite.ai

Scite.ai is an AI research platform built around Smart Citations, a proprietary classification system that analyzes how scientific papers are cited across academic literature. Rather than reporting raw citation counts, Smart Citations classify each citation statement as supporting, contrasting, or mentioning the cited work. This gives researchers immediate signal on whether findings have been validated, contested, or simply referenced in later literature. The platform indexes 1.2 billion+ citation statements and 280M+ full-text articles through licensing agreements with major publishers. Features include the Scite Assistant (AI Q&A grounded in scientific literature), full-text search, custom dashboards for tracking papers and authors, citation alerts, journal and institutional analytics, and an official MCP server for AI agent integration. Founded in Brooklyn in 2018 and funded in part by the National Science Foundation and the National Institute on Drug Abuse (NIH), Scite.ai was acquired by Research Solutions, Inc. in November 2023 and now operates as part of its research intelligence portfolio. The platform serves academic researchers, pharmaceutical scientists, science journalists, and research-intensive organizations globally. Scite.ai offers meaningful integration depth across several channels. Its official MCP server is published in the Model Context Protocol registry (endpoint api.scite.ai/mcp, documented at scite.ai/mcp) and lets AI agents including Claude and ChatGPT access Smart Citation analysis programmatically across 280M+ full-text articles. A Zotero plugin brings citation analysis directly into the leading open-source reference manager. A browser extension surfaces Smart Citation context while browsing publisher sites. Enterprise-tier API access supports institutional and programmatic integration, and publisher data agreements supply full-text PDFs and XML directly from major publishers for citation classification. Named integration gaps: no Mendeley native connector, no Overleaf or LaTeX integration for manuscript workflows, and API access is gated to the Enterprise tier rather than the individual plans. Scite pricing is publicly listed across three individual and team tiers, with custom Enterprise pricing. Basic ($20/month, or $12/month billed annually at $144/year): unlimited Scite Assistant queries, full-text search, Smart Citation Reports, citation alerts, Claude and ChatGPT connectors, 250 MCP credits per month, and 1,000-paper Collections. Pro ($50/month, or $40/month annually at $480/year): everything in Basic plus 2,500 MCP credits per month, 10,000-paper Collections, and Patents, clinical trials, and grants datasets. Team ($100/month, or $80/month annually at $960/year): everything in Pro plus 2 included users, 2,500 MCP credits per user, shared Collections and alerts, and user analytics, with additional seats at $40/month each. Enterprise (custom pricing, contact sales): everything in Team plus unlimited users, pooled usage, Organization API access, regulatory and safety datasets (FAERS, MAUDE, 510(k), MHRA), SAML/SSO, and a dedicated customer success manager. A 7-day free trial is available; there is no permanent free tier. Scite.ai is not suited for researchers who need comprehensive literature synthesis beyond citation context. The platform classifies citations but does not autonomously run multi-round adaptive literature searches, generate full annotated research reports, or extract structured data across many studies. For depth-first citation-trail search with full annotated reports, Undermind ($16/mo billed annually) covers that workflow. For AI-generated paper summaries and structured data extraction, Elicit ($49/mo billed annually) provides a dedicated systematic review pipeline. For open-domain research beyond peer-reviewed literature, Gemini Deep Research (free tier; $19.99/mo Pro) or ChatGPT Deep Research ($20/mo) cover broader source territory. For basic citation lookup without AI analysis, Semantic Scholar (free) offers comparable coverage. Trustpilot reviewers note weaker performance on complex synthesis tasks compared with Perplexity AI and deep research tools; Scite.ai is strongest at citation-context classification, not general-purpose research generation. Scite.ai launched its official MCP server at scite.ai/mcp in 2025, enabling AI agents to access Smart Citation analysis programmatically and positioning Scite as a citation-aware data layer for agentic research systems. In 2026 the plans were restructured into four tiers (Basic, Pro, Team, and Enterprise), with MCP access metered in monthly credits (250 on Basic, 2,500 on Pro). A 7-day free trial remains available with no permanent free tier as of Q3 2026. Trustpilot shows 223 reviews at a 3.8 TrustScore; G2 shows 27 reviews at 4.7/5. Negative Trustpilot feedback documents AI hallucinations in the Scite Assistant, including fabricated citations with DOI links, broken BibTeX export, UX quality issues, and customer support response-time concerns. Coverage is strongest for indexed publisher literature; preprint and niche-field coverage have gaps noted by users. NSF and NIH funding provides institutional stability alongside commercial operations.