SciTrue
How SciTrue works

Evidence you can see

SciTrue is four tools for working with real evidence: verify a claim, search the literature, check the citations in a text, and ground a draft in real sources. Every result traces back to a source you can open, analyze, and chat with.

01 · Transparent attribution

Every argument traced to real evidence

Ask a question or input a claim, and SciTrue retrieves real studies and shows, for each one, how strongly it supports your claim: the study type (often a critical variable for the field), the journal's quality, the exact paragraph it relied on with the key sentence highlighted, and the validity conditions that say when the finding holds and when it doesn't.

Evidence strengthStrong
Study type: Meta-analysisVery High: Q1 journal✓ Peer-reviewed
Coffee consumption and health: umbrella review of meta-analyses
Exact paragraph
Across multiple outcomes, the umbrella review found that moderate coffee intake was associated with a significantly lower risk of type 2 diabetes in a dose-response manner, with the largest benefit at three to four cups per day.
Validity conditions
  • Holds for moderate intake (three to four cups per day).
  • Observational cohorts, so an association, not proven cause.
02 · Search papers

Ask in plain language, get real papers back

Describe what you're looking for the way you'd say it out loud, and SciTrue returns real studies as full evidence cards: study type, journal credibility, evidence strength, the exact paragraph with the key sentence highlighted, the validity conditions, and Ask this paper. Narrow with filters for year range, field, publication type, open access, and minimum citations, then refine by study type. It's built for literature discovery and systematic reviews: fast, filterable, and every result traceable to its source.

Is there a paper on intermittent fasting and HbA1c in older adults?
Year: 2018–2025Field: MedicineType: Clinical trialOpen access
Study type: RCTHigh: Q1 journalEvidence strength
Time-restricted eating and glycemic control in adults over 60: a randomized trial
Over 12 weeks, time-restricted eating lowered HbA1c by 0.3 percentage points versus usual care in older adults with prediabetes.
💬 Ask this paperFor systematic reviews
03 · Verify your text

Catch fake citations and overclaims, and replace them with real ones

Did you ask a question to AI and get a confident answer? Even the best models can invent fake evidence and citations. SciTrue resolves each citation to the real paper and checks whether it backs the claim, flagging fabricated references, misattributions, and overclaims. When a citation is fake or off-target, it finds real papers on the claim so you can swap them in. Hover any highlight for the source quote and the reasoning.

Check citations

Resolves every citation to the real paper and checks whether it backs the claim, flagging fabricated references, misattributions, and overclaims.

Ground your draft

Finds real papers for a claim or an uncited draft and returns a citation-ready version, with the sources and references included.

Coffee is one of the most studied beverages in the world. Large reviews have found that moderate coffee intake is associated with a lower risk of type 2 diabetes (Ding et al., 2014). Looking at long-term data, some writers conclude that the more coffee you drink, the longer you live (Poole et al., 2017). Other posts get the direction wrong, warning that drinking coffee regularly raises your long-term blood pressure (Xie et al., 2018). A few even claim that drinking coffee significantly raises your risk of cancer (Harvard Health, 2023). One blog even cites the diabetes study for the idea that coffee improves athletic endurance (Ding et al., 2014), and adds that coffee protects long-term brain health (Bennett, 2021).

SupportedPartially supportedOverstatedContradictedNot addressedCitation not foundhover a highlight →
04 · Ask the paper

Question any source, grounded in its text

Did an argument catch your interest, or do you want to know more detail about how and why a study was actually done? Open the paper and ask. Answers come only from that article's own text, with the supporting passage highlighted, so you are never handed an ungrounded guess.

💬 Ask this paper
💬 Asking: Coffee consumption and health (BMJ, 2017)
How did they measure the link?
It pools observational cohorts and reports a dose-response association with lower risk, strongest at three to four cups a day. Being observational, it shows association, not proven cause.
From the paper
We found that coffee consumption was associated with a lower risk of type 2 diabetes in a dose-response manner.
05 · Use it your way

Different ways to use SciTrue

The same verification engine, wherever you work: on the pages you browse, inside your AI assistant, or built right into your own product.

Browser extension

Check the science as you browse

SciTrue rides along on YouTube and the news, surfacing the claims worth checking, or highlight any claim on any page to check it yourself.

Add to Chrome
MCP connector

Use it in your AI assistant

Connect SciTrue to Claude, ChatGPT, or any MCP client to search the literature, verify claims, and check citations against real research, right in the chat.

See the MCP connector
REST API

Build it into your product

Call the SciTrue API from your own software to search for papers or verify claims, and get back the same evidence cards, validity conditions, and references you see in the app.

Read the API docs