For Science Journalists & Researchers
Surgisphere. Amyloid fraud. The Stanford president. Every science scandal started with a study someone trusted without checking. Verify any paper in 2 minutes — before your story runs.
13,000+
Papers retracted in 2023
1 in 7
May contain fraudulent data
2 min
To verify any paper
Verify research from the world's leading journals
Real papers. Real scores. Click to preview.
Three steps from press release to verified story.
Got a press release with a study link? A PDF from a PR flack? Paste the identifier or upload the file directly. Works on any published paper.
Statistical power calculations, methodology benchmarking, conflict detection, and citation quality assessment — the kind of vetting a specialist scientist would do, in minutes.
Comprehensive report with scores, red flags, citation-ready language, and stronger alternative studies. Know whether to run the story, add caveats, or kill it entirely.
Investigative-Grade Analysis
ChatGPT gives you summaries. We give you ammunition.
Specific funding amounts, undisclosed conflicts, endpoint switching, statistical inadequacies — the concrete evidence you need to know if a study is worth reporting on, or worth avoiding.
Every report includes post-hoc power calculations, sample size adequacy assessment, and field benchmarks. Know instantly if a study was too small to detect its own claimed effect — the kind of flaw that makes a story fall apart after publication.
Example: "N=47 is underpowered for this endpoint — results may not replicate."
See exactly how a study's design compares to the gold standard for its research question. We rank methodology on evidence hierarchy so you can tell readers how seriously to take a finding — not just that a study "suggests" something.
"RCT vs observational vs case report — where does this study land on the evidence ladder?"
Stop wondering how to hedge a finding. Every report includes ready-to-use language for your story — plus alternative papers that are methodologically stronger, so you can report with appropriate confidence.
"While Smith et al. (2024) suggests X, the study's observational design limits causal claims..."
We surface the specifics that matter: exact funding amounts, institutional affiliations, protocol deviations, endpoint changes, p-hacking indicators, and cherry-picked outcome reporting. The details that expose hype — and protect your byline.
"Primary endpoint changed post-registration; secondary endpoint promoted to primary."
A press release lands in your inbox. The study looks solid — published journal, university PR behind it, catchy finding. Your editor wants it by 4pm.
That's exactly how it happened with Surgisphere. The Lancet published a hydroxychloroquine study that shut down WHO trials worldwide. It was fabricated. Journals retracted it. Journalists who reported it had to run corrections. The amyloid fraud that misdirected Alzheimer's research for 16 years. The Stanford president who resigned over manipulated data. Every one started as a study someone trusted on deadline.
13,000+ papers were retracted in 2023 alone — and Retraction Watch tracks over 48,000 total, many still being cited years later. 1 in 7 papers may contain fraudulent data.
A correction costs more than a subscription. Verify before you publish.
Paper Integrity scans for the red flags that slip past reviewers — and past reporters on deadline. Funding conflicts, methodology gaps, buried limitations, and credibility red flags.
Who funded the research? Do the authors have financial ties to outcomes? We surface what's disclosed — and flag what's missing. The pharma funding buried on page 14.
Were participants chosen to favor certain results? We examine recruitment methods and sample representativeness — the fine print that makes "promising results" meaningless.
Is the study design appropriate? Are the statistical methods sound? We assess the scientific foundation so you know if the headline matches the evidence.
What results were emphasized? What was buried in the appendix? We look at the full picture, not just the abstract — which is all most PR flacks want you to read.
Get comprehensive analysis with numerical scores, methodology assessment, and actionable insights — fast enough to make your deadline.
Download reports or generate social posts to share analysis with editors, colleagues, or your audience. Show your work, protect your credibility.
Ready to verify your next story?
Our Story
For science journalists, researchers, and anyone who needs to know if a study is real.
Origin Story
It started the way most ideas do — with frustration. And it came from both sides of the lab bench.
Every week, another link would land in the group chat. A friend sharing a study about supplements. A family member forwarding a paper on vaccine efficacy. A colleague citing research on environmental toxins. And every time, the same question: “Is this legit?”
Our founder spent years in biology research — the kind where you read methodology sections for breakfast and argue about sample sizes over coffee. So naturally, he became the guy people asked. Can you look at this study? Who funded it? Is this journal even reputable?
He'd spend thirty minutes pulling apart a single paper. Checking the funding disclosures. Cross-referencing the authors' publication history. Evaluating whether the conclusions actually matched the data. And then he'd do it again the next day for someone else.
But it wasn't just the non-scientists struggling. Science journalists were fighting their own version of the same battle — on deadline, with a press release in hand, trying to figure out if the study behind it was worth reporting. Researchers were spending hours vetting citations. The volume was crushing. Hundreds of thousands of papers published every month, and the manual process of evaluating each one hadn't scaled in decades.
The tools had changed. The rigor required to use them hadn't.
The Stakes
Science has a credibility problem — and it's not because the science itself is failing. It's because the infrastructure for evaluating it hasn't kept pace with the output.
For science journalists, the pressure is structural. You cover beats that require trusting studies you don't have time to fully vet. A PR agency sends a press release with a credentialed author, a real journal, a compelling finding. The story writes itself — until it doesn't. Surgisphere published in The Lancet and NEJM simultaneously before anyone checked if the data was real. The Alzheimer's amyloid fraud misdirected $40 billion in research for 16 years. Marc Tessier-Lavigne resigned the Stanford presidency over manipulated data.
For researchers, the challenge is volume. Funding conflicts buried in supplementary materials. Retracted papers that continue accumulating citations years after withdrawal. Reproducibility crises across entire fields.
That gap between published and trustworthy is where bad decisions get made — in newsrooms, in clinics, and at kitchen tables.
2.8M+
Papers published annually on PubMed alone
13,000+
Retractions in 2023 — and climbing fast
1 in 7
Papers may contain fraudulent or manipulated data
The Solution
Paper Integrity was built to close that gap — for journalists, researchers, and non-researchers alike.
We use AI to analyze scientific publications the way a seasoned peer reviewer would: examining funding sources, author conflicts of interest, methodological rigor, statistical validity, journal reputation, citation quality, and reproducibility indicators. The output scales to the reader.
Verify before you publish. Check funding conflicts, methodology red flags, and retraction status before a study becomes your story — or your correction. The Chrome extension works right in your browser, on deadline.
Triage incoming allegations faster. Screen submissions for methodology red flags, undisclosed conflicts, and statistical anomalies before committing to a full investigation. A rapid first-pass that never gets fatigued.
A rapid triage layer for literature review. Flag potential issues in funding, methodology, or citation networks before you invest hours in a deep read. A second set of trained eyes on every paper you cite.
Paper Integrity is proudly built in America by a small, independent team — no big tech affiliations, no institutional conflicts, no agenda. Just a commitment to making research transparency accessible to everyone.
$29/month is expense-report friendly. One avoided correction pays for years of coverage.
Browse real analyses
For individual journalists
For teams & desks
Everything journalists and researchers need to know before trusting an AI with their verification workflow.
Yes — with appropriate framing. Paper Integrity is a screening and analysis tool that surfaces red flags for further investigation. Use our findings to ask better questions, identify experts to contact, and decide whether a study merits the caveats you'll add to your story. We provide reporter-ready language in every report to help you describe findings accurately. As with any tool, your editorial judgment remains the final word.
Paper Integrity is a screening tool, not a replacement for peer review. We identify red flags — funding conflicts, methodological gaps, statistical issues — that warrant closer examination. Think of us as a thorough first-pass that catches what hurried readers miss. The final judgment is always yours.
Yes. Your submissions are encrypted in transit and at rest. We process papers solely to generate your report — we don't use them to train AI models, share them with third parties, or retain them longer than necessary. You can delete your data anytime.
We analyze multiple dimensions: funding source disclosure, author conflicts of interest, sample selection methods, statistical rigor, outcome reporting completeness, and citation context. Each dimension is scored and weighted based on established research integrity frameworks. Full methodology details are included in every report.
Plagiarism checkers find copied text. Paper Integrity analyzes research quality — whether a study's conclusions are supported by its methods, whether conflicts are disclosed, whether the statistics hold up. We're checking if the science is sound, not if the words are original.
Absolutely. Most users analyze papers they're reporting on or citing in their own work. If you can legally access a paper, you can analyze it. Paste the DOI from a press release, upload a PDF, or paste a PubMed link.
You can — but you'll get generic answers. Paper Integrity cross-references Retraction Watch, funding disclosure databases, and citation networks. We check if a paper has been retracted, flagged, or criticized in ways ChatGPT can't see. Plus, our structured methodology catches the same red flags every time — no prompt engineering required, and no hallucinated citations.