Research paper plagiarism checker.
Self-check your manuscript before the journal runs iThenticate. 700M+ scholarly works (OpenAlex + CORE + arXiv + PMC + Crossref) indexed. Co-author allowlist via ORCID. Preprint-overlap detection. Flat $79/mo Premium for unlimited revisions vs iThenticate's $125-per-document credits — matters across the typical 3-5 round peer-review process.
Four sections. Four different thresholds.
A journal manuscript isn't uniform prose. The Methods section legitimately shares boilerplate with prior papers. The Discussion legitimately cites peer work heavily. The classifier knows which section it's reading and applies the right thresholds.
Lit-review density expected
Introductions cite heavily — that's the section's job. The classifier accepts PARA-CITED density without inflating the headline score; flags UNCITED-PARA at full weight. A 1,200-word introduction with 30 citations is normal; one with 3 citations and 600 words of suspicious paraphrase is the problem.
Boilerplate is legitimate
Methods sections share standard text across the lab's prior papers. “Cells were maintained in DMEM supplemented with 10% FBS at 37°C in 5% CO2” appears identically across hundreds of your lab's papers — and should. Add lab heritage text to allowlist; tags as LAB-CITE, excluded. Different from accidental copying of someone else's protocol.
Original by definition
Your results section should be entirely original. Any non-trivial match here is a flag worth investigating: maybe accidental boilerplate copy from a template, maybe a sentence drift from your prior paper. The classifier weights matches in Results higher than matches in Introduction; the threshold for what counts as concerning is lower.
Peer-citation density expected
Discussions reference peer work to contextualize findings. Like Introduction — PARA-CITED density is normal, UNCITED-PARA is the issue. The classifier accepts comparing your results to e.g. (Smith et al. 2023) extensively as long as the citation is present and the comparison is novel synthesis, not text reuse.
Three checks across the manuscript lifecycle.
Pre-submission is one check. Revision rounds are another. Final acceptance is a third. The economics of iterative checking matter — flat-rate Premium means revisions don't burn per-document credits.
01 · Pre-submission self-check
Run the full manuscript before clicking submit. Catches the preprint-overlap (you posted to arXiv last month — journal's own iThenticate WILL flag that overlap; better to declare it in your cover letter than have the editor surprised). Confirms the methods boilerplate is properly allowlisted. Verifies the discussion citations are present. The PDF report attaches to your cover letter as evidence of due diligence.
02 · Revision-round checks (R1, R2, R3)
Reviewers ask for new analyses, expanded discussion, additional citations. Each revision is a substantive change to the manuscript. Re-check before re-submitting. iThenticate's $125-per-document model gets prohibitive at three revision rounds; flat-rate Premium means iteration is free. Most papers go through 2-4 rounds before acceptance.
03 · Final acceptance verification
Some journals run iThenticate at acceptance (not submission). Final-version check catches any text drift introduced during revision rounds, particularly if you used Track Changes carelessly and accepted suggestions that didn't change meaning but did change phrasing. Engine commit stamped on the report makes the final check reproducible if questioned post-publication.
Six things multi-author manuscripts need that single-author tools don't.
Multi-author papers introduce coordination problems single-author thesis tools don't have: whose prior work counts as self-citation? what's the preprint overlap policy? how do co-author CRediT contributions get tagged?
01 · YOUR OWN PRIOR PAPERS
ORCID + DOI allowlist. Your own published journal papers tag as SELF-CITE — important for follow-up studies where you legitimately build on your prior work. Tag includes the journal name; some journals have specific reuse policies (Cell Press allows methods reuse, Nature requires full disclosure of overlap > 10%).
02 · CO-AUTHOR PRIOR PAPERS
Add collaborator ORCIDs to allowlist. Tags as CO-SELF-CITE. Useful for multi-PI labs where your senior author has 200+ prior papers, many sharing boilerplate with your current manuscript. The classifier separates legitimate lab-corpus reuse from accidentally lifting from a collaborator's paper outside your lab.
03 · LAB BOILERPLATE
PI-folder shared corpus: standard methods sections, statistical procedures, ethics statements that propagate across the lab's publications. Tag as LAB-CITE, excluded from headline. The IRB-approved boilerplate (e.g. “This study was approved by the institutional review board, protocol #...”) doesn't need to be rewritten for every paper.
04 · PREPRINT OVERLAP
Your arXiv / bioRxiv / medRxiv / SSRN preprint of the same manuscript. Add preprint DOI to allowlist; tags as PREPRINT-SELF. Critical because journal-side iThenticate WILL flag preprint overlap if not declared. Most journals (Nature, Cell, NEJM, PLOS) accept preprints with explicit disclosure in cover letter; the classifier surfaces the matched intervals for declaration.
05 · CONFERENCE PROCEEDINGS
Add ACL / ACM / IEEE / NeurIPS conference DOIs to allowlist. Different journals have different policies on conference-to-journal extension — some allow with attribution, some require 30%+ new content. Tagged CONF-SELF-CITE; the report surfaces the matched intervals so you can verify the journal's specific reuse policy.
06 · DUPLICATE SUBMISSION
Add your own past manuscripts (in-press, under-review, drafts shared with collaborators) to a tenant-only folder. The classifier surfaces overlap with these as DUPLICATE-SUBMIT-RISK — the unethical case where the same paper is submitted to multiple journals simultaneously. Catches it before the cross-journal editorial network does.
What the report looks like before you click submit.
IMRaD-section-aware classification with co-author + preprint allowlist active. Headline number reflects what your journal's iThenticate will actually flag; the rest is exclusions with reasons.
Why iThenticate's $125-per-document model gets prohibitive across a typical peer-review cycle.
Solo researcher, 3 revision rounds
Pre-submission + R1 + R2 + R3 + acceptance check = 5 checks. iThenticate: 5 × $125 = $625. Noplag Premium: $79/mo × 4 months = $316. Saves ~50% per paper. Reuse the saved Premium across multiple papers and the ratio improves further.
Most common scenario · cumulative savings compound5-author paper, lab corpus reuse
iThenticate per-document doesn't have co-author allowlist; co-author's prior work flags as suspect overlap, requires manual disclosure to the editor. Noplag's CO-SELF-CITE handling means matched intervals tag correctly; cover letter writes itself.
Multi-PI labs benefit most · explicit collaborator handlingYour target journal is a Crossref Similarity Check member
If your journal is a CSC member (most major publishers are), they get iThenticate at member-discounted rates. The journal's check happens regardless of what you ran first. Noplag's role is the iterative self-check; the journal's iThenticate is the final gate. Use both.
Journal-side iThenticate happens regardlessThe classifier writes the preprint-disclosure language for you.
Most major journals (Nature, Cell, NEJM, PLOS, Science) explicitly allow preprints with disclosure in the cover letter. The mechanics: include a sentence like “This manuscript has been previously deposited as a preprint at bioRxiv (DOI: 10.xxxx/2025.03.18). The submitted version is substantively unchanged from the preprint, with the addition of [new figures / additional analyses / expanded discussion of X].” The classifier generates this language automatically based on the matched intervals between your preprint and your submitted manuscript — listing the specific sections that changed, the percentage of new text, and the preprint DOI. Copy-paste into your cover letter. Same applies to conference-proceedings reuse (“This work was previously presented at ACL 2025 [DOI]; the journal version expands the analysis with...”). Saves the editorial back-and-forth where the editor asks “is this previously published?” — you've already answered.
Read the cover-letter guideWhat corresponding authors actually ask.
- Does this replace iThenticate at journal submission?
- No — the journal will run iThenticate (or Crossref Similarity Check) at submission regardless of what you ran first. Noplag is the iterative self-check before that submission, plus the revision-round checks during peer review. Most authors use both: Noplag during writing + revisions; iThenticate is what the editor's screen shows.
- How does preprint disclosure work?
- Add your preprint DOI (arXiv / bioRxiv / medRxiv / SSRN) to the allowlist. Matched intervals tag as PREPRINT-SELF and surface separately from the headline score. The classifier generates cover-letter language listing the specific sections that changed, the overlap percentage, and the preprint DOI — copy-paste into your submission. Most major journals (Nature, Cell, NEJM, PLOS, Science) explicitly allow preprints with this disclosure.
- What about co-author papers — do they count as my own?
- Add collaborator ORCIDs to the allowlist. Matched intervals against your co-authors' prior papers tag as CO-SELF-CITE (excluded from headline; surfaced separately for cover-letter declaration). Useful for multi-PI labs where the senior author has 200+ prior papers, many sharing methods boilerplate with your manuscript.
- Does it work for journals in my language (non-English)?
- Detection is calibrated for English, Spanish, Portuguese, Polish, and Ukrainian at launch, with more languages rolling out. Detection quality is benchmarked on PAN-PC-11 and published at /docs/developers/benchmarks; per-language evaluation is on the roadmap. Useful for European-language journals (Revista Médica de Chile in Spanish, Acta Universitatis Carolinae in Czech) and Lusophone journals (Brazilian medical journals).
- What's the IMRaD-aware classification actually doing?
- The classifier detects which section it's reading (Introduction / Methods / Results / Discussion) and applies appropriate thresholds. Methods-section boilerplate from your lab's prior papers tags as LAB-CITE (excluded). Results-section matches weight higher because Results should be original by definition. Discussion citations of peer work tag as PARA-CITED low-weight when attribution is present. The headline number reflects section-appropriate flagging.
- Will my manuscript be used to train AI?
- No. Manuscripts are fingerprinted (winnowing hashes, non-reversible) and stored in your tenant-isolated database. The text is purged after 30 days unless you opt into retention. Detection models are pre-trained on public corpora; we don't add manuscripts to training. Explicit in the DPA — your co-authors' institutions can require it.
- How will the AI detector handle a paper drafted with co-author input?
- AI detection is coming soon — it's on the roadmap, not live at launch. When it rolls out: per-paragraph AI-likelihood (Binoculars, Hans et al. 2024). The verdict will be paragraph-level — useful for multi-author papers where different sections were drafted by different authors. An ESL-adjusted band is planned for international collaborations where some co-authors are non-native English. Tag pattern: a paper with mixed AI-likely paragraphs across sections is a different signal than a paper where one specific co-author's contributions read AI-likely.
- Can I self-host inside my institution's network?
- Yes — Apache 2.0 + Docker. Common at European research institutions with strict GDPR-residency requirements for sensitive primary research (clinical data, patient records, classified material). Pair with your institutional Postgres + Redis; no submission traffic leaves your VPC. Useful for manuscripts under embargo or protective order.
- What about cumulative-PhD-style papers reused in dissertations?
- Common scenario: a journal paper becomes a dissertation chapter. The classifier handles both directions: at journal submission, the dissertation context doesn't matter. At dissertation submission later, add the journal DOI to your dissertation allowlist; matched intervals tag as SELF-CITE. Most institutions explicitly allow this with a preface note about which chapters are based on which published papers.
- How does pricing work for a department or research group?
- Premium at $79/mo per user, or Enterprise contracts for departments with 10+ active authors. Enterprise adds: shared lab corpus (PI-folder boilerplate visible to all group members), SSO/SAML integration with your institutional auth, EU residency endpoint, dedicated support SLA. Most research groups end up on Enterprise after the first quarter; the lab-corpus feature is what they upgrade for.
Run the manuscript. Generate the cover-letter language. Submit clean.
Drop the manuscript .docx or .pdf into Noplag. See IMRaD-aware classification, co-author + preprint allowlist active, cover-letter language generated. Free tier covers 2,500 words for a section sanity check; Premium at $79/mo handles 50,000 words and unlimited revision rounds.