HomeAI Detector › For PhD Students

AI Detector for PhD candidates, from literature review to final viva chapter.

Pre-scan dissertation chapters before your supervisor reads them, and pre-scan journal manuscripts before iThenticate or Crossref Similarity Check sees them. Sentence-level highlights show exactly which lines react AI, with perplexity and burstiness per sentence so you can edit specific prose instead of rewriting whole chapters. Calibrated for citation-dense lit review and theoretical framework writing. GDPR aware, no training on dissertation drafts. Free to try. No card.

Start free, no card See pricing
Institutional Pro at $13.99/mo GDPR aware 90-day chapter history
Who it is for

Built for dissertation chapters and journal articles.

A PhD is multi-deliverable and multi-year. Literature review, candidacy or qualifying exam, prospectus, dissertation chapters, journal submissions, and conference papers all carry different stakes and different supervisor visibility. The pre-submission scan that makes sense for one stage rarely fits the next without adjustment.

The PhD writing stack runs from the first scoping memo to the bound thesis. Pre-scanning fits every layer because the institutional check at the end, whether Turnitin AI on your LMS or iThenticate on your journal manuscript, runs the same kind of classifier on every chapter, every paper, every grant.

Literature review and theoretical framework

Citation-dense paraphrasing by design. Heavy formal register. Naturally flags higher than other chapters because the patterns overlap with machine paraphrase. The sentence-level highlights, not the headline number, are the diagnostic that matters here. Free tier covers single-section scans up to 5,000 characters. Pro at $19.99 a month, or $14.99 a month on yearly, unlocks 10,000 character pastes and unlimited scans for the iteration weeks.

Methods, results, and discussion chapters

Methods carries precise procedural claims that the AI rewriter should leave alone. Results is mostly tables and statistical reporting. Discussion is where your original voice should dominate and your Authenticity Score targets are highest. Pre-scanning catches AI-shaped phrasing in discussion before a supervisor questions it.

Journal submissions and conference papers

Journal flags during peer review can mean desk rejection or a revise-and-resubmit with explanation. Conference papers drafted under deadline pressure benefit from the same workflow. The 90-day Pro history keeps every submission scan retrievable, useful when an editor or programme chair asks about a manuscript you sent three weeks ago.

Chapter-by-chapter workflow

Scan by chapter, calibrated for each section type.

A typical PhD chapter runs 8,000 to 15,000 words. Pro caps each scan at about 1,600 words. The split into sections is not a limitation. It matches how supervisors read drafts and how internal examiners reread them. Calibration targets differ by section.

Literature Review: aim 55 to 70

Citation-dense scholarly paraphrase reads structurally close to AI paraphrase. Expect lower scores than other chapters and rely on the sentence map. Scattered yellows in formally structured prose are usually register, not residue. Red sentences clustered in one paragraph are worth a rewrite pass with the Light AI rewriter.

Methods: aim 70 plus

Methods is procedural and should read tight. Aim for 70 plus and use the AI rewriter sparingly. Statistical reporting, sample descriptions, instrument descriptions, and analytical procedures carry precise claims. If the AI rewriter is needed, use Light mode only so meaning is preserved sentence by sentence.

Results: aim 70 plus

Results prose is short and table-adjacent. The scan reads tables poorly, so paste the narrative paragraphs and skip the table cells. Expect higher scores here than in lit review, since the prose is your own short sentences carrying numbers.

Discussion and Conclusion: aim 75 plus

This is where your original voice should dominate, and where supervisors look hardest. A discussion that reads AI-shaped is the section that draws committee questions. Aim 75 plus, and rewrite any red sentence rather than rewriting it. The conclusion follows the same standard.

Supervisor, candidacy, viva

Sentence-level evidence for supervisor and committee conversations.

A supervisor conversation about AI-shaped writing goes better when you can point at specific sentences instead of defending a percentage. The same evidence trail works for candidacy or qualifying exams, the prospectus signoff, and the final viva or defence.

Pre-scan before every chapter handoff

Run TextSight on each section before emailing the draft to your supervisor. The 90-day history keeps every scan retrievable. When your supervisor flags a paragraph, you can open the matching scan and see which sentences the classifier reacted to and which it passed. Often it is the same paragraph.

Share the PDF report if a question comes up

The PDF export stores the input text, the Authenticity Score, the sentence-level colour map, the timestamp, and the classifier version. If your supervisor or doctoral committee asks about a section, the PDF is something specific to discuss instead of a vague defence. That shifts the conversation from "did you use AI" to "let us look at these three sentences".

Candidacy or qualifying exam responses

Written candidacy responses are multi-hour committee essays where AI-shaped prose is the worst place for a false positive. Pre-scan the response while drafting, especially literature-heavy sections, and aim for sentence-level cleanliness rather than chasing a single number. The same approach fits the prospectus, which blends lit review and methodology in formal register that flags higher than discussion prose.

Audit trail across the dissertation cycle

A PhD spans years. Submitting a finished chapter without knowing how a year-old draft scored is a needless risk. Pro history plus PDF archives give you the audit trail across the cycle. Useful when an internal examiner asks about a draft from eight months ago that you have since revised twice, or when a viva question lands on a paragraph you wrote in year one.

Disclose proactively, not reactively

Most institutions now require AI disclosure in some form. The defensible pattern is early specific disclosure: tell your supervisor which tools you used for which task, share TextSight scan reports as you produce them, and let the committee see the audit trail rather than discover it during defence. The tool exists to give you that evidence.

Plans & pricing

Pricing for PhD candidates and post-docs.

Verified .edu, .ac.uk, .ac.in, and .edu.au emails get the institutional discount on Pro automatically at signup. Pro is $19.99 a month standard, $14.99 a month on yearly, and $13.99 a month with institutional verification. Full details on the pricing page.

Free
$0/forever

 

Sample a single chapter section or conference abstract.
  • 3 scans / day
  • 5,000 chars per scan
  • Sentence-level highlights
  • 2 lifetime AI rewriter uses
Start free
Starter
$7.49/month

Billed $89.88/year — Save $30

For a candidate writing a section a week between research cycles.
  • 20 scans / day
  • 20,000 AI rewriter words/mo
  • Chrome extension
  • Email support
Get Starter
Business
$29.99/month

Billed $359.88/year — Save $120

For doctoral cohorts, lab groups, and writing centres.
  • 100,000 AI rewriter words/mo
  • 5 team seats, shared history
  • Audit log, REST API
  • White-label PDFs
Get Business

.edu, .ac.uk, .ac.in, .edu.au, .edu.sg, .edu.ph, .edu.vn, and .ac.ke emails get Pro at $13.99/mo. The discount applies automatically at signup. View full pricing →

Pre-Turnitin and pre-journal

Defensible before iThenticate, Crossref, and Turnitin AI.

Two institutional checks bracket your PhD writing. Turnitin runs on chapter submissions through your university LMS. iThenticate and Crossref Similarity Check run on journal manuscripts before peer review. Pre-scan both with the same workflow so neither one surprises you.

Step 1: draft normally

Write in your usual editor: Word, Docs, Overleaf for LaTeX, or your university LMS. Using ChatGPT for an outline, a lit review brainstorm, or to break writer's block is the realistic 2026 default. Write the prose itself in your own voice from your own notes and your own reading.

Step 2: paste and scan by section

Open app.textsight.ai, paste each chapter section, and scan. Free tier handles 5,000 characters in one paste. Pro handles 10,000. For a 12,000-word chapter, split by section and scan each in turn. The scan returns in about thirty seconds with an Authenticity Score and a sentence-by-sentence colour map.

Step 3: edit by section type

Discussion and conclusion: above 75, submit; below 75, rewrite reds in your voice. Lit review and theoretical framework: above 55, the sentence map is the signal, not the score. Methods and results: above 70, submit; rewrite reds rather than rewriting precise procedural claims.

Step 4: archive the scan, then submit

Save the PDF report or rely on the 90-day Pro history. Hand the chapter to your supervisor or the manuscript to the journal. If a question comes up later, the saved scan plus the version-controlled draft gives you the audit trail. A typical chapter section round-trips in about fifteen minutes; a 6,000-word journal manuscript in about forty.

What you see in a scan

Perplexity, burstiness, paragraph cards, and 90-day history.

A single percentage is not a fix path on a chapter draft. The TextSight result panel shows which sentences reacted and why, with paragraph-level rollups for longer chapter sections, so you can edit the specific lines instead of rewriting the whole submission.

Sentence-level highlights

Every sentence is colour-coded by its own AI-likeness score. Red sentences clustered in one paragraph are a stronger signal than scattered yellows. Scattered yellows in otherwise structured lit review prose often just mean the writing is formally taught. You read the pattern, not just the headline number.

Perplexity, read-only on Pro

Perplexity is how predictable your word choices are to a language model. Low perplexity reads AI-like. The score is shown per sentence on Pro, which is the diagnostic context you need to decide whether a flag is real AI residue or just an unusually well-rehearsed theoretical framework intro.

Burstiness, read-only on Pro

Burstiness is how much your sentence length and structure vary across the section. ChatGPT defaults to uniform medium-length sentences. Real scholarly writing has bursty rhythm: one short sentence, one long, one fragment, one parenthetical. Low burstiness across an entire chapter is the classic AI fingerprint and the one supervisors learn to spot first.

Paragraph cards for longer drafts

For dissertation chapters and journal manuscripts, paragraph-level rollups identify which sections drift AI-shaped and which stay clean. Useful when you have a 9,000-character methods chapter and need to know which two paragraphs to revisit rather than rereading the whole section.

90-day history on Pro

Every scan is retrievable for 90 days. For a writer iterating across a 6-month chapter cycle, that means every clean section scan and every revision is on record. PDF export lets you save longer-term archives chapter by chapter, beyond the 90-day window.

Journal pre-submission awareness

Aware of Nature, Science, Elsevier, Wiley, IEEE, Springer, and Taylor & Francis screening.

Major scholarly publishers now run AI detection on incoming submissions as a default screen, in some cases alongside iThenticate and Crossref Similarity Check. A flag during peer review can mean desk rejection, revise-and-resubmit, or in rare cases a misconduct inquiry. Pre-scanning every manuscript is the defensible posture in 2026.

What publishers screen for

Most large publishers, including Nature, Science, Elsevier, Wiley, IEEE, Springer, and Taylor & Francis, now require some form of AI-use disclosure in submission policies. Several of them also run automated AI detection on incoming files. The screen is not always a hard block. It is usually a flag that asks the handling editor to decide.

iThenticate and Crossref Similarity

iThenticate runs the same kind of similarity check Turnitin runs on student submissions, against the broader scholarly corpus through Crossref Similarity Check. AI detection is being layered on top. Pre-scanning your manuscript with TextSight gives you the sentence-level map before iThenticate generates the report your handling editor reads.

What to do if a submission is flagged

Run the TextSight scan, save the PDF, and reach the handling editor with specific edits. Flagged sentences with a saved revision history are easier to defend than a general denial. The 90-day Pro history makes this concrete: you have a record of what you scanned, when, and what changed between submission and revision.

Honest disclosure of AI tools

The most defensible journal posture is honest disclosure of which AI tools were used for which tasks, in the methods or acknowledgments. ChatGPT for outlining, Grammarly for proofreading, a translation tool for ESL drafting. The TextSight scan is the evidence that the final prose is yours, not a way around the disclosure.

Conference and grants

Applied to NSF, NIH, ERC, and Wellcome grant prose.

Conference deadlines and grant submissions cluster, and they cluster late. The same scan workflow that fits a chapter section also fits an 8-page conference paper and a multi-section research proposal. Free tier covers a single section; Pro lifts the cap so a full grant narrative clears in a sitting.

Conference papers under deadline

Rapid AI-assisted drafting under deadline pressure is the realistic 2026 pattern for many conference submissions. The honest workflow is to draft fast, then run TextSight on each section, then rewrite reds in your own voice and re-scan. An 8-page paper clears Pro's daily budget without rationing. Free tier requires scanning section by section.

Grant prose for NSF, NIH, ERC, Wellcome

NSF, NIH, ERC, and Wellcome funding bodies look hard at the prose in proposals, especially specific aims, significance, and innovation sections. AI-shaped grant prose is becoming a reviewer red flag even where formal policy is silent. Pre-scan each section, rewrite reds, then submit. The PDF report is useful if a programme officer asks downstream.

Abstract and impact sections

Conference abstracts and impact statements are short and high-stakes. They are also where stock phrasings creep in fastest. The 5,000-character free tier covers a typical abstract in one scan, the 10,000-character Pro cap covers most impact sections in one paste. The sentence map matters more here than the headline score.

Co-author and lab-group workflow

Business tier adds 5 team seats, shared history, and a REST API. Useful for lab PIs running multiple PhD students through the same grant or paper pipeline, or for doctoral cohorts coordinating around a thesis-by-publication track. Audit logs and white-label PDFs come standard on Business.

Your dissertation stays yours

Privacy first: no training on submitted text, GDPR aware.

PhD drafts are unpublished scholarly work. They are also protected by GDPR in the EU and UK, by FERPA in the US, by the DPDP Act in India, and by local equivalents elsewhere. TextSight is designed to honour those rules out of the box, not as a paid setting you have to find.

No training on submitted text

Chapter drafts, journal manuscripts, conference papers, and grant prose submitted for scanning are never used to train the classifier or any other model. This is a contract clause, not a configuration toggle. It applies on the free tier the same way it applies on Pro and Business.

No account required on free

The free tier needs no email, no account, no identity. For candidates worried about institutional disclosure or pre-publication leakage, this matters. You can scan a draft chapter without TextSight ever knowing who you are or which institution you attend.

Your supervisor and examiner do not see your scans

Scan history is private to your account. We do not share scan data with universities, supervisors, doctoral committees, examination boards, journal editors, conference programme chairs, Turnitin, iThenticate, or any third party. Your scans are not part of any institutional or editorial record.

Deletion on request, DPA on Business

Any scan can be deleted from your history. On Pro you can delete individual records. Data retention is bound to your settings, and a standard DPA is available on Business and Enterprise tiers for university writing centres, lab groups, and doctoral cohorts.

FAQ

PhD candidates frequently ask.

Can TextSight scan a full PhD dissertation?
Pro caps each scan at 10,000 characters, about 1,600 words. A full PhD dissertation must be scanned chapter by chapter, and within long chapters, section by section. The split actually matches how supervisors and examiners read drafts, so the granularity is useful. The 90-day Pro history keeps every section scan retrievable for the cycle of revisions before submission to the examination board.
How does TextSight handle citation-dense literature review prose?
Lit reviews are scholarly paraphrasing by design and they flag heavier than other chapters because the patterns overlap with machine paraphrase. Use the sentence-level highlights to see which specific sentences reacted. Scattered yellows across a structured lit review are usually formal register, not AI residue. Clusters of red sentences in one paragraph are worth rewriting. Aim for an Authenticity Score in the 55 to 70 band on lit review and use the sentence map for the rest.
Will my supervisor or examiner see my TextSight scans?
No. Scans are private to your account. The free tier does not require email or identity. Paid tier scan history is visible only to you. We do not share data with your university, supervisor, doctoral committee, examination board, journal editor, Turnitin, iThenticate, or any third party. Your dissertation drafts and journal submissions are not part of any external record we control.
How does TextSight sit beside Turnitin and iThenticate?
Turnitin runs on your university LMS for chapter submissions and viva drafts. iThenticate and Crossref Similarity Check run on journal manuscripts before peer review. TextSight is the pre-flight scan for both. It correlates within 5 to 10 percentage points of Turnitin AI in internal testing across 3,100 graduate submissions. Treat the institutional check as the source of truth, and TextSight as the workshop you run on every chapter and every manuscript before sending.
What is the .edu discount for PhD candidates?
Verified .edu, .ac.uk, .ac.in, .edu.au, .edu.sg, .edu.ph, .edu.vn, and .ac.ke emails get Pro at $13.99 a month instead of $19.99, applied automatically at signup. The discount is the same for PhD candidates, master's students, and undergraduates. Pro includes unlimited scans, a 10,000 character cap per scan, 90-day scan history, file upload, and the integrated AI rewriter. If your domain is not auto-recognised, contact support with your student ID for manual verification within 24 hours.
Does TextSight train its model on submitted dissertation chapters?
No. Text submitted for scanning is never used to train the classifier or any other model. This is a contract clause, not a configuration toggle. Data retention is bound to your history settings, deletion on request is supported, and our privacy practices are GDPR aware in the EU and UK, FERPA aware in the US, DPDP aligned in India, and aligned with local equivalents elsewhere. Dissertation chapter drafts are treated identically to undergraduate essays.
What about false positives on PhD writing?
False positives on dense academic prose are real, especially in literature reviews, theoretical frameworks, and methods sections where formal register and structured paraphrasing dominate. The TextSight classifier was tuned against graduate submissions from Indian, Filipino, Nigerian, Chinese, Vietnamese, and Kenyan student populations, and false positive rates are roughly 40 percent lower than for US-only competitors. Sentence-level highlights are the diagnostic tool that matters, not the headline score on its own.
What if my candidacy or defense committee questions a draft?
Run the scan and save the PDF report. The report stores the input text, the Authenticity Score, the sentence-level flags, the timestamp, and the classifier version. That is the format a doctoral committee or research integrity panel actually wants to see, and it gives you something specific to discuss instead of a vague defence. The 90-day history means you can also show how an earlier draft scored and how revisions moved the score before submission.
Related

More for PhD candidates.

Pre-scan a chapter. Submit clean. Defend confidently.

Free to try. No card. Institutional Pro at $13.99/mo for verified .edu, .ac.uk, .ac.in, and .edu.au emails.

Start free, no card See pricing
GDPR aware · No training on dissertation drafts · Sentence-level highlights · Institutional Pro discount