Scite.ai (Scite Inc. / Research Solutions, Chicago 2018)

Status: Active (founded 2018 by Josh Nicholson et al.; acquired by Research Solutions NASDAQ: RSSS Dec 2023; NSF + NIH SBIR funded; 1.6B+ Smart Citations, 300M+ sources, ~2M users; scite.ai reachable 2026-08, no discontinuation). Category ai-productivity-data; pipeline_stage=none.

1. Introduction

Scite.ai is a citation-intent verification layer for scientific literature, not a general-purpose research chatbot. Its core unit is the Smart Citation: for any paper it shows not just how many times it was cited, but whether each citing paper supports, contrasts, or merely mentions the original claim, with the exact sentence context pulled from full text. Layered on top are Scite Assistant (grounded Q&A over the corpus), Reference Check (retraction/editorial-notice scan on a manuscript’s reference list), citation alerts, dashboards, Zotero + browser-extension badges, and a 2025 MCP server so external AI agents can query citation sentiment programmatically. It replaces “cited-by count” with “cited-how”, which is why systematic reviewers, grant writers, and fact-checkers reach for it before trusting a number from Google Scholar.

2. Core Features

Layer	Capability	Notes
Smart Citations	Classify each citation as supporting / contrasting / mentioning; show snippet + section context	1.6B+ classified citations, 300M+ sources; deep-learning classifier, confidence scores shown
Scite Assistant	Natural-language Q&A grounded in full-text corpus, answers carry verifiable citation statements	LLM-based; not autonomous synthesis, hallucination risk on complex asks
Reference Check	Upload manuscript PDF → scan every ref for retraction status, editorial notices, support/contrast balance	Used by journals/preprint screeners
Advanced Search	Query citation statements not just metadata; filter by support/contrast, journal, year, section	Full-text licensed deals (Wiley, BMJ, Cambridge) give paywalled reach
Dashboards + Collections	Track a DOI set; alert on new supporting/contrasting citations or retractions	Basic 1,000 / Pro 10,000 papers per collection
Integrations	Chrome extension (Scholar/PubMed badges), Zotero plugin, ChatGPT/Claude connectors, REST API, MCP server (250–2,500 credits/mo by plan)	Only Pro+ gets self-service API key; MCP = agent-facing
Evidence datasets	Patents, clinical trials, grants, drugs/devices/safety (Pro+)	Need dataset scopes on API

3. Pricing (2026-08, USD)

Tier	Monthly	Annual (eff/mo)	MCP credits	Key limits	Notes
Free	$0	—	0	Public Papers/Tallies endpoints, no Assistant, no Smart Citation reports behind login wall	Browse-only; 7-day trial auto-enrolls to chosen paid plan
Basic	$20/mo	$12/mo ($144/yr)	250/mo	Assistant unlimited, full-text search, 1,000-paper collections, ChatGPT/Claude integrations	Individual researchers; no REST API key
Pro	$50/mo	$40/mo ($480/yr)	2,500/mo	+ self-service API key, 10,000-paper collections, patents/grants/clinical trials in search	Power users / API evaluators
Team	$50/user/mo	$40/user/mo ($480/user/yr)	2,500/user/mo	Pro per seat, pooled org collections, central billing	≤20 seats typical self-serve
Enterprise	Custom	Custom	Pooled	SSO/SAML, managed credentials, extra Evidence scopes, dedicated CSM	Research Solutions sales-led

pricing_type=subscription-mcp-credit; starting_price=$0; no permanent free tier (7-day trial only, auto-renew unless cancelled); student/academic discounts via institutional referral.

4. Access Type

Field	Value
access_type	web-browser-extension-zotero-mcp-api-no-desktop-app
access_display	🌐 scite.ai (browser) · Chrome extension (Scholar/PubMed) · Zotero plugin · iOS/Android web-responsive · MCP server (scite.ai/mcp) · REST API (Pro+) · no offline desktop client

5. Reviews (2026)

Source	Rating	Sample	Signal
G2	4.7/5	~27	Ease-of-use, Smart Citation accuracy, literature-review speed praised; small n
Trustpilot	3.8–4.1/5	~220	Strong tool, but assistant hallucinations (fabricated DOI), broken BibTeX export, cancel/billing friction, support latency
Capterra	sparse	<10	No robust sample; mirror G2 pros
Academic eval (Bakker 2023 etc.)	—	—	Supporting/contrasting precision decent, recall lower esp. for contrasting; misclassification on nuanced frames

Verdict: best-in-class for “is this finding still standing?” — beats Scholar/Semantic Scholar on intent. Tradeoffs = Assistant not trusted for autonomous synthesis (human must verify), niche/preprint/non-EN coverage thinner than biomed, MCP/credit metering surprises trial→paid migrants, $20 floor stings students without library access.

6. Best For / Not For

Best For	Not For
Systematic reviewers / meta-reviewers checking if a cited result held up	General-purpose research generation (Perplexity / ChatGPT Deep Research broader)
Grant writers & postdocs stress-testing their reference list	Pure discovery engine for “what’s new in field” (Semantic Scholar free)
Journal editorial screens / retraction checks (Reference Check)	Non-English / humanities / very recent preprint-heavy fields (coverage gaps)
Pharma / regulatory writers needing patent/clinical-trial citation context	Users expecting zero hallucination from Assistant (needs human pass)
AI-agent builders wanting citation-sentiment MCP tool	Students without institutional subsidy (price friction at $20/mo)

7. Competitors

Lane	Tool	Starts	Scite’s edge / gap
Citation-intent	Consensus	$10.99/mo	Consensus = abstract consensus meter; Scite = sentence-level support/contrast
Discovery (free)	Semantic Scholar	$0	S2 free, counts+influential; no intent classification
Agentic lit review	Elicit	~$49/mo	Elicit extracts structured data across studies; Scite classifies citations
Deep research	Perplexity, ChatGPT Deep Research	$20/mo	Broader web; not citation-intent-grounded
Citation graph	Litmaps, Connected Papers	freemium	Visual citation networks; no support/contrast text
Systematic review	Rayyan	freemium	Screening/dedup; Scite complements, doesn’t replace
Academic LLM	Scispace (Typeset)	freemium	PDF Q&A; Scite stronger on citation sentiment DB