Scite AI Just Changed How I Trust Research - And It’s Free (Scite AI Tutorial)
Based on Andy Stapleton's video on YouTube. If you like this content, support the original creators by watching, liking and subscribing to their content.
Start with Scite AI’s free browser extension and Zotero plugin because they deliver the core trust signals without subscription cost.
Briefing
Scite AI’s biggest value for researchers comes from its free browser extension and Zotero plugin, which surface citation-quality signals directly while searching and managing papers. Instead of treating citations as a single “how many” metric, the extension adds a compact set of numbers under Google Scholar results—publications citing the work, supporting citation statements, mentioning citation statements, and contrasting citation statements. Hovering and clicking those badges routes users to Scite’s paper page, where the meaning becomes clear: supporting statements indicate agreement, while contrasting statements flag where other authors cite the work while disagreeing or raising problems. That contrast-focused view helps readers develop a “sixth sense” while browsing—spotting papers that are widely cited for positive reasons versus those that attract disagreement, and even identifying potential research gaps when a work is cited mainly in contested contexts.
The tutorial emphasizes that these signals are most useful when they’re embedded in everyday workflows. The browser extension places the Scite indicators under search results, and the same information can appear in a side toolbar when users open the publisher page. For reference management, the Zotero plugin mirrors the same citation-statement numbers inside Zotero, making it easier to evaluate sources without leaving the library. Installation is described as straightforward: download an XPI file and drag-and-drop it into Zotero’s plugin interface. Once installed, adding references triggers a Scite sidebar with the key numbers, and a “site report” link takes users back to Scite’s paper page for deeper context.
Paid offerings exist, but the tutorial frames them as less compelling than the free tools. The “assistant” chat interface can generate reference-backed answers (with options like specifying evidence sources such as citation statements and abstracts, and ranking results by relevance). Users can also switch the underlying GPT model (the example uses Claude) and receive an AI summary alongside the cited references. Still, the workflow is described as somewhat uneven compared with other research assistants, and the free tier is portrayed as restrictive—limited search history and only a small number of searches—pushing some users toward a $20/month plan.
Beyond chat, Scite’s dashboard and aggregation features are presented as convenient but not essential, especially for researchers who already have the citation signals in Zotero. The most notable paid feature is “reference checks,” positioned for end-of-draft verification: it flags whether cited references have editorial concerns such as corrections, using Scite’s citation-statement data to show which references are contested. Alerts, metrics, and profile features round out the paid ecosystem, but the overall takeaway is pragmatic: start with the free extension and Zotero plugin because they deliver the core trust-and-verification value at no cost, and only consider paying if additional features—like reference checks—fit a specific publishing workflow.
Cornell Notes
Scite AI’s strongest research utility comes from its free browser extension and Zotero plugin, which display citation-quality signals under search results and inside reference libraries. The key numbers separate citations into supporting, mentioning, and contrasting citation statements, letting researchers gauge not just how often a paper is cited but whether the citing community agrees or disputes it. The tutorial argues that contrasting statements are especially useful for spotting contested work and potential research gaps. Paid tools add chat-based assistance and dashboards, but the free tier is limited and some paid features are described as less standout than competing tools. “Reference checks” is highlighted as the most valuable paid capability for sanity-checking references near submission time.
What do Scite’s citation badges mean, and why does the supporting/contrasting split matter more than total citations?
How does the browser extension change literature searching in practice?
Why is the Zotero plugin described as a workflow upgrade rather than a standalone feature?
What does Scite’s paid assistant do, and what limitations are mentioned?
Which paid feature is singled out as most useful for researchers preparing to submit work?
Review Questions
- When evaluating two highly cited papers, which Scite metric(s) would you prioritize to estimate community agreement, and which would you treat as a red flag?
- How would you use the browser extension versus the Zotero plugin differently during the research process?
- What is the purpose of “reference checks,” and how does it relate to the contrast between supporting and contrasting citation statements?
Key Points
- 1
Start with Scite AI’s free browser extension and Zotero plugin because they deliver the core trust signals without subscription cost.
- 2
Use supporting citation statements as evidence of community agreement, not just total citation counts.
- 3
Treat contrasting citation statements as a warning indicator that other authors dispute or challenge the work.
- 4
Install the Zotero plugin by adding the downloaded XPI file via Zotero’s plugin tools, then review Scite metrics in the Zotero sidebar.
- 5
Paid chat assistance can generate reference-backed answers, but the tutorial frames it as less standout than competing research tools.
- 6
“Reference checks” is the most clearly useful paid feature for pre-submission verification of cited sources and editorial concerns.
- 7
Scite dashboards and profile features are described as helpful snapshots, but not essential if citation signals already live in Zotero.