CliniPaperAI manuscript review for medical research papers
CliniPaper reviews a clinical manuscript the way a demanding peer reviewer would, before you submit. It recomputes every number across the abstract, tables and results, verifies each reference against PubMed and Crossref, checks the methods and the reporting guideline for your design, and returns the problems first, each with an exact fix you can paste.
Review your manuscript
Describe your study or paste your text. CliniPaper starts the work and checks every number and reference.
What does an AI manuscript review check?
| Area | What CliniPaper checks |
|---|---|
| Numbers | Recomputes every n/N, percentage, subgroup total and crude OR/RR/RD from your counts, and checks that the same number matches across abstract, results, tables and figure legends. |
| References | Confirms authors, title, journal, year, volume, pages and DOI of each reference against PubMed and Crossref, checks that the reference supports the sentence citing it, and flags anything it cannot verify. |
| Methods | Design label, selection and missing data, confounding, variables on the causal path, collider bias, interaction claims, power for null findings and time-trend analysis. |
| Reporting | Checks the manuscript against the reporting guideline for your design (STROBE, CONSORT, PRISMA, CARE…): flow chart, baseline table, n/N everywhere, ethics and AI-use disclosure. |
| Journal fit | Abstract structure, word limits, reference style and required statements from your target journal's author instructions. |
The reporting check follows the EQUATOR guideline for your design. See which reporting guideline your study needs, the STROBE checklist for observational studies and the CONSORT checklist for trials.
How does it work?
- Attach your draft. Upload the manuscript (Word or PDF) and any summary tables, and name the target journal if you have one. Summary tables only: no identifiable patient data.
- CliniPaper reviews it. It reads the paper like a reviewer: recomputes the numbers, verifies the references, checks the methods and the reporting guideline.
- Fix what matters first. Major issues come first, each with why it matters and the exact fix (analysis or wording, ready to paste). Then minor issues and ideas that strengthen the paper.
What does a review look like?
Each problem comes with the reason it matters and a fix. An example with synthetic data:
Major issue 1: the abstract and Table 2 disagree. The abstract reports 312 of 4,200 deliveries (7.4%) with the composite outcome; Table 2 reports 298 (7.1%). The same count must appear everywhere, and the composite cannot be smaller than any of its components in Table 3.
Fix, ready to paste: “The composite adverse outcome occurred in [XX] of 4,200 deliveries ([XX]%).” Placeholder until you confirm the count from your dataset.
References are marked verified or not found, so nothing fabricated reaches the journal. Learn how to format them in Vancouver style or AMA style.
What happens after the review?
Ask CliniPaper to apply the fixes, restructure the paper to your target journal’s author instructions, draft the cover letter and the AI disclosure statement, and later answer the reviewers point by point.
Frequently asked questions
Is it allowed to use AI to review my manuscript before submission?
Usually yes. ICMJE and most publishers allow authors to use AI tools if they disclose it and stay responsible for the content; a few journals restrict AI-generated text, so check your target journal’s policy. What journals prohibit is reviewers uploading other authors’ unpublished manuscripts to AI tools. CliniPaper also drafts the disclosure statement for you.
How is this different from ChatGPT?
General chatbots can invent references and rarely check arithmetic. CliniPaper verifies every reference against PubMed and Crossref, recomputes your numbers, and marks anything it could not verify instead of guessing.
Is my unpublished manuscript used to train AI models?
No. Your files are never used to train models, ours or our providers’, and you can delete a paper and its files at any time. Do not upload identifiable patient data; CliniPaper does not need it.
Does it replace a statistician or my co-authors?
No. It finds the problems a reviewer would raise and proposes fixes; you and your co-authors decide. For complex models, use the review to focus the questions you take to your statistician.
Which papers does it work for?
Original clinical research (cohort, case-control, cross-sectional, trials), systematic reviews, case reports, theses being turned into papers and revisions with reviewer comments.
How much does a review cost?
You can start free. A full review uses credits from your plan; you see the cost before long tasks run. See pricing on the homepage.
Start with the paper you already have in progress
Describe your study or paste your text. CliniPaper starts the work and checks every number and reference.
CliniPaper is a writing and review assistant. Authors remain responsible for their content and must disclose AI use according to journal policy.