Scite.ai
Scite.ai answers research questions from the scientific literature itself and shows whether later work supported or contradicted a finding.
If your work hinges on whether a claim still holds up, Scite answers a question Google Scholar can't: who contradicted this. Smart Citations classify each reference as supporting or contrasting a specific claim, and the Assistant ties every answer to a source sentence you can check. The free Connect tier (25 MCP credits/month inside ChatGPT or Claude) is a genuine no-cost way to test that signal before paying. Paid entry is Basic at $14/mo billed yearly (30% fall-sale rate, promo ends September 30th, 2026) for full Assistant and Search access; Pro at $35/mo billed yearly adds API access and patent, clinical-trial, and grant datasets. If you only need quick lookups, PubMed is free and faster.
Verified 7d ago · liveness 97/100 · cite: rightaichoice.com/tools/scite-ai
- Systematic reviewers and meta-analysts who need to know if a finding was later supported or contradicted
- PhD students and postdocs placing a paper in the context of the wider literature
- Grant writers tracing an idea from funded proposal to published finding
- Librarians and research support staff running evidence-based reference services
- Quick lookups where Google Scholar or PubMed is faster and free
- Anyone working primarily in non-English literature, where coverage is limited
- Undergraduates writing term papers who won't use citation classification
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Scite if your research question only needs a fast free lookup on PubMed or Google Scholar rather than a verdict on whether a finding was later supported or contradicted.
Going past your monthly MCP credit allotment stops your agent workflow — Pro raises the ceiling to 2,500 credits/month, but heavy automation can still exhaust it.
The free Connect tier costs nothing and covers 25 MCP credits/month inside ChatGPT or Claude. Basic at $14/mo billed yearly fits individual researchers who need full Assistant and Search. Pro at $35/mo billed yearly suits teams needing API access plus patent, trial, and grant data — cheaper than most enterprise literature databases, but pricier than free tools like PubMed. Enterprise pricing is scoped via sales@scite.ai.
In short
Scite.ai — Scite.ai answers research questions from the scientific literature itself and shows whether later work supported or contradicted a finding. Best for Systematic reviewers and meta-analysts who need to know if a finding was later supported or contradicted, PhD students and postdocs placing a paper in the context of the wider literature, Grant writers tracing an idea from funded proposal to published finding. Free to start; paid plans from $14/mo.
What's new in Scite.ai
Checked 7 days agoAcross the latest 5 updates: 1 feature update, 3 changelog entries and 1 news mention.
Scite MCP adds citation graph tool for AI agents
Scite MCP now includes a citation graph tool that lets AI agents like ChatGPT and Claude explore how papers connect through citations.
August 2026 Release Notes
August release added easier connections to AI assistants, more generous trial limits, simplified plan selection, improved library import, and research monitoring.
Best ChatGPT Plugins for Research 2026
Scite was listed among ChatGPT research plugins alongside Zotero, Article Galaxy, Wolfram, BioRender, and custom MCP apps.
July 2026 Release Notes
July updates focused on staying on top of relevant research and giving more visibility and control over Scite use across AI tools.
June 2026 Release Notes
June release made Collections fully manageable inside the MCP, added AI-powered suggestions across the platform, and sped up setup for small teams.
Viability Score
How well maintained and how widely used is Scite.ai? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- Smart Citations classify each citation as supporting, contrasting, or mentioning a claim
- Scite Assistant answers research questions grounded in specific sentences from real papers
- One-click jump from any AI answer to the underlying source sentence
- Full-Text Search across 300M+ articles, preprints, books, patents, and datasets
- Reads across 317M+ full-text scholarly articles via 44+ publisher agreements
- Patent Search covering 90M+ patent families
- Clinical trials search across 719K+ records
- Grants search across 3.5M+ funded projects
- Collections gather papers, monitor updates, and run scoped questions
- Collections manageable directly inside MCP (June 2026 release)
- MCP connector for Claude, ChatGPT, and other MCP-compatible assistants
- MCP citation graph tool lets AI agents explore how papers connect (September 2026)
- Zotero plugin surfaces citation data inside your reference library
- Browser extension displays citation badges anywhere
- Reference Check verifies and corrects manuscript citations
About Scite.ai
Scite.ai is a research platform built on what it calls Smart Citations: instead of just counting how many times a paper was cited, it classifies each citation as supporting, contrasting, or merely mentioning a claim, method, journal, author, institution, or funder. That classification rests on 1.6B+ citation statements drawn from 317M+ full-text scholarly articles, licensed through direct agreements with Wiley, SAGE, and 44+ publishers plus the open-access corpus. The AI Assistant uses those same evidence signals, and every answer links back to the specific sentence in the specific paper, checkable in a click. Full-Text Search goes past abstracts; Collections let you gather, monitor, and run scoped questions against your own paper set — and since June 2026 Collections are manageable directly inside Scite's MCP connector. The database spans journals into 90M+ patent families, 719K+ clinical trials, 3.5M+ grants, preprints, and datasets, so an idea can be traced from funded proposal to published finding to commercial filing. Scite runs inside Claude, ChatGPT, and any MCP-compatible assistant, plus a Zotero plugin, browser extension, and API access on Pro. It's for systematic reviewers, PhD candidates, grant writers, librarians, and IP teams — people who can't afford a wrong citation. Against Semantic Scholar, Elicit, and plain Google Scholar, the differentiator is explicit citation classification backed by licensed full text rather than abstract-only inference.
Behind the Verdict
Scite's case rests on one thing competitors can't easily copy: licensed full text plus a citation-classification model trained on 1.6B+ citation statements. Google Scholar tells you a paper was cited; Scite tells you whether the citing paper backed the claim up or knocked it down. For a meta-analyst deciding whether to include a study, or a grant writer checking whether a mechanism was later refuted, that's the difference between a defensible reference list and a liability. The verification chain is the second genuinely useful piece — every Assistant answer links to the specific sentence in the specific paper, so you're not re-checking the tool's work from scratch. And Scite has pushed its distribution well past its own app: since the June 2026 release, Collections are manageable inside the MCP connector; the July 2026 release added visibility and control over Scite use across AI tools; August 2026 brought easier AI assistant connections, more generous trial limits, and simplified plan selection; and the September 2026 addition of a citation graph tool lets agents like ChatGPT and Claude explore how papers connect through citations. That last one matters because it lets you keep your existing writing workflow and still get grounded citations instead of model hallucinations. Weaknesses are honest ones. The free Connect tier gives 25 MCP credits/month with no Assistant or Search, and Basic caps at 250 — heavy agent users will burn through that. Full-text reach depends on publisher and institutional agreements, so coverage tracks the licensed and open-access corpus. Non-English literature remains thin. Access to patents, clinical trials, and grants sits behind Pro at $35/mo billed yearly, and API work needs Pro too. Scite is now part of Research Solutions, which brings stability and institutional reach but also means the product roadmap is set by a public company. Bottom line: Scite is worth its price for citation-integrity workflows and isn't worth it as a general-purpose search engine. If you need to prove a claim is still standing, nothing else does quite what Smart Citations do.
Researching Scite.ai? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Scite.ai actually fits — and what changes day-one when you adopt it.
You're screening a candidate study for a meta-analysis and need to know whether its headline claim was later supported or contradicted. You open its Smart Citation report, filter to contrasting citations, and read the source sentences directly.
Outcome: You exclude the study with a documented reason instead of discovering the contradiction during peer review, and your exclusion rationale cites the contrasting paper.
You draft a paragraph in ChatGPT and want every claim tied to a real paper. With the Scite MCP connected to ChatGPT, the Assistant pulls supporting and contrasting citations, and collections of your key papers are managed directly inside the same conversation.
Outcome: Each sentence in your draft carries a checkable citation, and you catch one claim where the literature is contested before your advisor does.
You need to trace a published finding into commercial filings. On Pro you run Full-Text Search across journals and patents, then check whether the underlying paper was supported or challenged before citing it in a freedom-to-operate memo.
Outcome: You map the idea from funded proposal to published finding to patent family in one platform rather than stitching together three databases.
Use Cases
- Evaluate a paper's credibility before citing it in your own work.
- Get a quick, evidence-backed literature review on a research question.
- Check a manuscript's references for retracted or contradicted studies.
- Track how a finding has been supported or contradicted over time.
- Monitor a research field for new supporting or contrasting citations.
- Search patents and clinical trials alongside academic literature.
- Verify claims in AI chat outputs using Scite MCP within Claude or ChatGPT.
- Let an AI agent trace how papers connect via the MCP citation graph tool.
Models Under the Hood
as of 2026-09-22
Limitations
- Scite's classification rests on its own deep learning model applied to 1.6B+ citation statements across 44+ publisher agreements; coverage is limited to that licensed and open-access corpus.
- MCP credits cap at 25/month on the free Connect tier — which has no Assistant or Search access — and at 250/month on Basic, so heavy agent users need Pro for 2,500.
- Full-text access depends on those publisher and institutional agreements, and answers are confined to indexed scholarly literature.
- Patent, clinical trial, and grant datasets sit behind Pro at $35/mo billed yearly.
as of 2026-09-30
Verification history
We have re-verified Scite.ai 19 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 19 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Scite.ai tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Free
$0/mo
Ideal for
Researcher who wants to test citation-grounded answers inside ChatGPT or Claude without paying
What this tier adds
Starting tier and free entry point: 25 MCP credits/month, no Assistant or Search access, no credit card required
Basic
$14/mo billed yearly (30% fall sale; promo ends September 30
Ideal for
Individual researcher or PhD student who needs full Assistant and Search across the Scite platform
What this tier adds
Adds Full Assistant and Search access, 250 MCP credits/month, article Smart Citation reports, Collections up to 1,000 papers, and citation alerts and dashboards
Pro
$35/mo billed yearly (30% fall sale; promo ends September 30
Ideal for
Team or analyst who needs programmatic access plus patent, clinical trial, and grant data
What this tier adds
Adds API access with a $60 one-time API credit (limited time), 2,500 MCP credits/month, Collections up to 10,000 papers, and patent, clinical trial, and grant datasets
Enterprise
Custom
Ideal for
Institution or organization needing custom seat terms and dataset scope
What this tier adds
Adds custom pricing and seat terms scoped directly with the sales team via sales@scite.ai
Where the pricing makes sense
The company stage and team size where Scite.ai's pricing actually pencils out — and where peers do it cheaper.
The free Connect tier costs nothing and covers 25 MCP credits/month inside ChatGPT or Claude. Basic at $14/mo billed yearly fits individual researchers who need full Assistant and Search. Pro at $35/mo billed yearly suits teams needing API access plus patent, trial, and grant data — cheaper than most enterprise literature databases, but pricier than free tools like PubMed. Enterprise pricing is scoped via sales@scite.ai.
Setup time & first value
How long it actually takes to get something useful out of Scite.ai — broken out by persona, not the marketing-page minute.
Free Connect tier: a few minutes — sign up, connect the Scite MCP in ChatGPT or Claude, and run your first grounded question. Basic: same-day — full Assistant and Search unlock immediately after signup, with a 7-day trial available first. Pro: same-day for API access, though teams building custom integrations should budget a day to wire up the API and review the $60 one-time API credit allowance.
Switching to or from Scite.ai
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From Google Scholar: export your key papers into a Scite Collection, then use Smart Citations to add the supporting/contrasting layer Scholar doesn't show.
- →From Zotero: install the Scite Zotero plugin so citation badges appear directly on items in your existing library.
- →From Semantic Scholar: recreate your saved paper lists as Scite Collections and run scoped Assistant questions against them.
- →From Elicit: import your reviewed papers into a Collection for ongoing monitoring instead of one-off literature passes.
- →From a manual spreadsheet: use Scite's improved library import (August 2026 release) to load your reference list as a Collection.
- ↗To Zotero: keep references in Zotero and retain the Scite plugin for citation badges if you leave the paid tiers.
- ↗To Google Scholar or PubMed: use these for free quick lookups once the citation-classification workflow isn't needed.
- ↗To Semantic Scholar: export the paper lists you built in Collections for a free citation-count view, without the supporting/contrasting classification.
- ↗To Elicit: move to a literature-review tool if your workflow centers on drafting a review rather than verifying individual claims.
Integrations
Resources & Guides
- Resourcescite.ai
AI for Research | Scite
Researchers around the world use Scite to better understand research, uncover debates, ensure they are citing reliable studies, and improve their writing.
- Quickstartscite.ai
AI for Research | Scite
Researchers around the world use Scite to better understand research, uncover debates, ensure they are citing reliable studies, and improve their writing.
- Resourcescite.ai
Scite Features
Research tools built on evidence, not guesswork. Every Scite feature is powered by 1.6B+ Smart Citations and 280M+ full-text articles — including paywalled sources other AI tools can't access.
Tutorials & Learning
YouTube returned 6 videos for “Scite.ai”, and we withheld 4: 4 could not be judged, because “Scite.ai” is a single word that other videos use for other things. Showing the 2 we can prove are about Scite.ai.
Official links
Tools that pair well with Scite.ai
Common stack mates teams adopt alongside Scite.ai, with the specific reason each pairing earns its keep.
openscience
Open-source, local-first AI workbench for the full scientific research lifecycle, with bundled scientific database tools and your choice of model provider.
WolframAlpha
Wolfram|Alpha answers computable questions with verified results computed from curated data and the Wolfram Language.
Typeform
AI form builder that turns every Typeform response into automated GTM and research workflows
Alternatives to Scite.ai
View allopenscience
Open-source, local-first AI workbench for the full scientific research lifecycle, with bundled scientific database tools and your choice of model provider.
WolframAlpha
Wolfram|Alpha answers computable questions with verified results computed from curated data and the Wolfram Language.
Frequently Asked Questions
Categories
Best-of guides
Topics
Used Scite.ai? Help shape our editorial sentiment research.