A service by Adeon Global

Originality, verified

Check any manuscript for similarity against the live web and for AI-generated writing, in one pass. Every scan produces a dated, shareable report you can attach to an editorial decision.

How it works

Three layers on every submission

Each layer answers a different question, and the report keeps them separate so you can see exactly where a score comes from.

Layer 1

Private corpus

Every manuscript we have already seen is fingerprinted and held locally. A resubmission, a recycled chapter or a paper shopped to two of your journals is caught here — instantly, and without touching an external service.

Layer 2

Web similarity

The text is scanned against live web sources. You get an overall similarity percentage, a ranked source list with per-source contribution, and the exact matched passages highlighted in the submitted text.

Layer 3

AI-writing detection

A separate model estimates how much of the text reads as machine-generated, sentence by sentence, and flags hidden-character tricks — zero-width spaces and homoglyph substitutions — used to defeat detectors.

Reading a report

One number, three bands

0 – 15% Within limits

Similarity at the level normal citation and common phrasing produce. No action needed.

15.1 – 25% Needs review

Worth an editor's eye. Usually long quoted passages, a shared methods section, or a boilerplate-heavy introduction.

Above 25% High similarity

Substantial overlap with existing sources. The report shows which sources and which passages, so the author can be asked directly.

Bands are configurable per institution — a review's threshold is an editorial policy decision, not a technical one.

Features

What ORIGEN does

Detection is provided by the Winston AI platform; ORIGEN adds the private corpus, the reporting and the workflow around it.

FeatureWhat it means in practice
AI content detection Detects whether submitted text is likely to have been generated by an AI model, on the same pass as the similarity check.
AI detection score A confidence score for how likely the text is to be AI-generated, shown against the threshold your institution has set rather than as a bare number.
Human vs AI classification Classifies the writing as human-written, AI-generated or somewhere between, so a borderline result is not forced into one box or the other.
Sentence-level analysis Scores every sentence individually and shades them in the report, so you can see where a draft reads as machine-written instead of judging it on one overall figure.
Detailed scan results Structured results — per-source contribution, matched passages with their positions, word counts and per-sentence scores — stored in full and rendered as a report, a PDF and a shareable link.
Multilingual support Submissions can be scanned in English, Hindi, Marathi, French, Spanish, German, Portuguese, Arabic and Chinese, or left on automatic detection. Accuracy varies by language and model.
Document & OCR handling PDF, DOC, DOCX and TXT are accepted. Text is read from the file directly where it has a text layer; a scanned document with none is passed to Winston, whose own document pipeline handles it. See the note below.
Plagiarism detection Similarity against live web sources, with a ranked source list, per-source contribution and every matched passage highlighted in the submitted text.

On scanned documents. ORIGEN does not run OCR itself. It reads the text layer of a file, and only when there is none does it hand the document to Winston for its own processing — which needs the upload to be reachable from the internet for the length of the scan. If you routinely submit scans or photographs of pages, tell us and we will confirm what your setup will do with them before you rely on it.

Around the detection

What AN-List adds

Your own private corpus

Every manuscript checked is fingerprinted and kept, so a resubmission or a recycled chapter is caught locally — instantly, and without spending a credit.

Exclusion filters

Leave out quotations, the bibliography, cited sources, short matches or individual sources. The score is recalculated from results already held, so changing your mind costs nothing.

Passage-level evidence

Every matched sequence is highlighted in the original text and numbered to the source it came from — not just a percentage.

Shareable reports

Each report has a permanent reference number and a signed link you can send to an author or reviewer without giving them an account.

Submissions stay private

Manuscripts are held in AN-List's own database, are not listed or indexed, and are never shared outside your institution's account. An uploaded file is deleted as soon as its scan finishes.

Built into your workflow

Run checks from your dashboard, or have your editorial office run them from the admin console with per-department monthly allowances.

Comparison

How ORIGEN compares with Turnitin

Turnitin is the tool most Indian institutions already know. Here is how the two differ in what they check, what you can do with a result, and who can actually use them.

  Turnitin ORIGEN
What it is built to do Text-similarity checking against a very large repository, with AI-writing detection added to the product since 2023. Similarity against live web sources and AI-writing detection in one pass, plus a private corpus that belongs to your institution.
AI-writing detection A document-level percentage with the sentences it attributes to AI highlighted, and guidance that the figure is not by itself a determination of misconduct. A score for the document and for every individual sentence, shaded across the text so you can see exactly where a draft reads as machine-written — with the flagging threshold set by your institution rather than fixed for you.
Plagiarism sources Academic publications, the web, and an archive of previously submitted student papers held by the vendor. Live web sources with passage-level matches, plus a private corpus of every manuscript your institution has checked — held in your own account, so a resubmitted paper or a recycled chapter is caught within your institution instantly and at no cost.
Where your submissions live Submitted work is added to the vendor's own repository. Manuscripts stay in your AN-List account. They are not published, indexed or shared outside your institution, and an uploaded file is deleted from the server as soon as its scan finishes.
Adjusting a result Quotations, bibliography and small matches can be excluded in the similarity report. Quotations, bibliography, cited sources, short matches and individual sources — recalculated from results already stored, so changing your mind never triggers a re-scan or costs anything further.
Who can use it Sold to institutions; students and staff normally reach it through their institution's own systems. Available to an institution, a single department or an individual — through an AN-List subscription or on its own. Browser-based, nothing to install.
Cost and visibility Institutional licence, arranged with the vendor. A monthly check allowance on your subscription, or unmetered for a department. Credits consumed and remaining are visible in your own admin console, so the cost of a scan is never a surprise.
The report Similarity report inside the institution's environment. A dated report with a permanent reference and a signed link that opens without an account, plus a PDF carrying the full highlighted text — so it can be attached to an editorial decision or sent to an author.

Built to show evidence, not just a number. Every detector on the market — Turnitin's included — can flag honest writing, and independent research has found this happens most often to students writing in a second language. ORIGEN is designed around that: it shows the matched passages and the sentence-level scores, so an examiner reads the text rather than a percentage, and a student can be asked about something specific. That is a better conversation, and a fairer one.

Comparison prepared August 2026 from Turnitin's own published documentation and guidance, and from independent published research. Turnitin is a trademark of Turnitin, LLC; AN-List and ORIGEN are not affiliated with, endorsed by, or a reseller of Turnitin. If you find anything here out of date or inaccurate, tell us and we will correct it.

Who it's for

What you get out of it

The same scan answers a different question depending on where you sit in the process.

Students

Find out where your draft stands before you submit it, while there is still time to fix it.

  • See exactly which sentences matched a source, so you can quote and cite them properly instead of guessing what upset the checker.
  • Learn the difference between honest overlap and a problem — a shared definition or a standard methods sentence is not misconduct, and the report shows why.
  • Catch a paraphrase that stayed too close to the original, the single most common reason good students get into trouble.
  • Get a dated report you can attach to a dissertation submission if your department asks for one.

Researchers

Protect work you have spent years on from an avoidable desk rejection.

  • Check a manuscript against the live web before it reaches an editor, including your own earlier papers — self-plagiarism is caught by the same scan.
  • Verify that a co-authored draft is clean when you are the corresponding author signing for all of it.
  • Screen a literature review, where dense citation makes a high raw percentage normal and the passage list is what actually matters.
  • Keep a record for a funder or ethics committee that asks how originality was assured.

Authors

Show that the work is yours, in a form someone else will accept.

  • A shareable report with a permanent reference number, which a publisher or a client can open without an account.
  • AI-writing detection alongside similarity — increasingly the first thing a commissioning editor asks about.
  • Evidence of authorship if your work is later copied, since every scan is dated and kept.
  • Book chapters, theses and reports all scan the same way; upload a PDF or DOCX and the text is read for you.

Academic scholars & supervisors

Judge a piece of work quickly, on evidence rather than on a number.

  • Sentence-level AI scores show where a draft reads as machine-written, which is a far better basis for a conversation than a single overall figure.
  • Hidden-character tricks — zero-width spaces, look-alike letters swapped in to defeat detectors — are flagged as findings in their own right.
  • Thresholds are set by your institution, because what counts as acceptable overlap is an academic judgement, not a technical default.
  • The private corpus catches a paper submitted twice, or recycled between two of your own courses or journals.

Editors & publishers

Screen submissions at volume without adding a step to your workflow.

  • Check any article already in your AN-List catalogue straight from the admin console — no upload needed where a PDF link or DOI is on file.
  • Per-source contribution and matched passages, so a rejection letter can point at something specific.
  • Citations are identified separately, so a well-referenced paper is not penalised for being well referenced.
  • Per-user monthly allowances, an audit log of every scan, and reports you can attach to an editorial decision.

Libraries & institutions

One originality service across departments, on the subscription you already hold.

  • Campus-wide access with no per-seat licence, and unmetered plans for heavy departments.
  • Submissions stay in your own account; uploaded files are deleted as soon as the scan finishes.
  • Usage is visible centrally, so a library can report on how the service is actually being used.
  • Nothing new to install — it sits alongside your eJournals, eBooks and case studies.

One caution worth stating plainly: no detector is infallible. A similarity score is evidence to be read, and an AI score is an estimate, not a confession. Both are reasons to look at the text — never a substitute for doing so.

Check your first manuscript now

Signed-in AN-List accounts get 10 checks a month. Subscriptions add a larger allowance, and institutional plans are unmetered.