Self-check coursework, master's papers, and dissertation chapters before they reach Turnitin or your supervisor. Sentence-level highlights show which lines read AI, with perplexity and burstiness signals so you can revise the prose in your own voice. Built with international and second-language writers in mind, since formal academic English is the writing most prone to false positives. Private to your account, never used to train our model. Free to try. No card.
University writing covers a wider arc than college does. Freshers writing their first 1,500-word essay, master's students submitting term papers under tight deadlines, and PhD candidates drafting full chapters all need the same pre-submission scan, but for different reasons.
The university stack runs from undergraduate seminars to master's coursework to doctoral chapter drafts. Pre-scanning fits every layer because the institutional report at the end is the same Turnitin AI check, regardless of whether you are in year one or the third year of a PhD.
Two to ten pages of structured argument across multiple modules. Free tier covers casual single-essay 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 weeks where you are submitting on rotation across three or four modules.
Heavier reading load, denser citation requirements, and supervisors who already know what AI-shaped prose looks like. The sentence-level highlights matter here because master's writing rewards specificity and a single AI-rewritten paragraph can be the one your supervisor questions. The 90-day Pro history is the safety net.
Chapter drafts that get scanned chapter by chapter as they come together. The 10,000-character cap forces you to scan in sections, which matches how supervisors actually read drafts. PDF export keeps a defensible record of which version of each chapter was scanned and when, useful when an examiner asks about a draft you submitted three weeks ago.
Turnitin still owns the institutional integrity record at most universities. TextSight is the private read you take on your own draft first, whether that draft is a first-year seminar paper or a methodology section bound for the examination board.
Word and Docs for coursework, Overleaf for a thesis written in LaTeX, or straight into Canvas or Blackboard. Using AI for an outline or a literature-search starting point is a question for your programme's policy. Write the prose itself from your own reading and notes.
Paste a completed essay, or one section of a longer chapter, into app.textsight.ai. The free tier takes 5,000 characters; Pro takes 10,000. A dissertation chapter is scanned section by section, which happens to match how supervisors read drafts anyway. Each scan returns an Authenticity Score and a sentence-by-sentence colour map.
Start with the red sentences, then the yellow clusters. Rewrite them as you would actually argue the point, add a specific source or finding, and vary the rhythm. For second-language writers especially, this is about confirming a flag is real residue rather than your formal register, not about chasing a number.
Run it once more to confirm the flagged lines moved, then submit through your LMS or send the section to your supervisor. A short essay round-trips in minutes; a dense chapter section takes a little longer because there is more to read.
Pro is $19.99 a month standard, or $14.99 a month on yearly billing. Full details on the pricing page.
Billed $89.88/year — Save $30
Billed $179.88/year — Save $60
Billed $359.88/year — Save $120
Yearly billing saves 25%. View full pricing →
A single percentage is not a fix path. 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.
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 prose often just mean you were taught to write formally. You read the pattern, not just the headline number.
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 literature review intro.
Burstiness is how much your sentence length and structure vary across the section. ChatGPT defaults to uniform medium-length sentences. Real human writing has bursty rhythm: one short sentence, one long, one fragment. Low burstiness across an entire chapter is the classic AI fingerprint and the one supervisors learn to spot first.
For dissertation chapters and master's papers, paragraph-level rollups identify which sections of a long draft drift AI-shaped and which stay clean. Useful when you have a 9,000-character chapter section and need to know which two paragraphs to revisit rather than rereading the whole thing.
Master's theses and PhD dissertations run far past the 10,000-character per-scan cap. The Pro workflow is to scan by section, archive each result, and treat the 90-day history as your draft audit trail before the examination board sees anything.
A typical PhD chapter runs 8,000 to 15,000 words. Pro caps each scan at about 1,600 words. Split a chapter into its natural sections, introduction, literature review, methodology, findings, discussion, conclusion, and scan each one in turn. The split matches how supervisors actually read drafts, so the granularity is useful, not punitive.
Every scan is retrievable for 90 days on Pro. For a writer iterating across a six-month dissertation cycle, that means every clean chapter scan and every revision is on record. PDF export lets you save longer-term archives chapter by chapter. When an examiner asks about a draft from three weeks ago, you have receipts.
Pre-scanning each chapter section before handing a draft to your supervisor catches AI-shaped phrasing before it reaches the person who will sign off your work. The conversation shifts from "did you use AI" to "this paragraph reads AI-shaped, let us discuss the underlying argument", which is the conversation you actually want.
Copy the rendered text into TextSight rather than the LaTeX source. The classifier reads the prose, not the markup, and citation commands or equation environments will throw off scores if pasted directly. The cleanest workflow is to compile, copy the body text from the rendered PDF or output, and scan that.
Native plugins are not shipped yet. Here is the honest 2026 picture of what works today and what is on the roadmap, so you can plan around it during your degree.
Draft inside Canvas, Blackboard, Brightspace, Moodle, or Google Classroom as you normally would. Before you click submit, copy the final text into TextSight at app.textsight.ai. Edit the flagged sentences in TextSight or back in your LMS, then submit the cleaned version. Round-trip is about six minutes for a typical undergrad essay.
Drag a DOCX, PDF, or TXT into TextSight if you wrote in Word, Docs, or exported from Overleaf. Pro accepts files up to 10,000 characters per scan and returns the same sentence-level result the paste-in workflow does. Useful when formatting matters and you do not want to lose it in copy-paste.
One-click scan from any web page including LMS submission views. Useful when you want to scan an assignment description, a peer review, or a paragraph from your own draft without leaving the tab. Available on Starter and above.
Canvas, Blackboard, Brightspace, Moodle, and Google Classroom plugins are on the 2026 roadmap. We are not promising dates while LMS plugin requirements keep changing each term. We would rather ship a working integration once than ship a thin wrapper that breaks every semester.
No single number proves a student used AI. Across higher education the expectation is that a detector reading is one input among several, weighed alongside drafts, sources, and a conversation. The case for self-checking sits on both sides of that.
A low Authenticity Score means your submission reads more AI-like to the classifier. It does not by itself mean you used AI, and it does not mean a supervisor will raise it. False positives are real, especially for second-language writers and for the dense, formal prose that postgraduate work rewards, where phrasing overlaps with AI defaults.
Keep the report. It records the text you scanned, the Authenticity Score, the sentence-level flags, and the timestamp. If a question ever arises, you can show a draft and revision trail rather than offering a flat denial, which is what an integrity panel is actually looking for.
Across a six-month dissertation cycle, AI-assisted drafting can creep further into the final text than you remember. If a section lands below 50 on prose you thought was your own, the honest fix is to rewrite those paragraphs from your notes, not to launder them through another tool.
When a student arrives having already self-checked and revised, a flagged paragraph becomes a discussion about the argument rather than an accusation. A detector reading is one input, weighed with drafts and sources, never a standalone verdict. That process protects honest writers, including those whose formal or second-language style is prone to false positives.
A self-check on unpublished research only works if it stays confidential. Your scans are yours, nothing reaches your university or supervisor, and your text is never used to train our model.
Essays, chapters, and theses you submit for scanning are never used to train our model or any other. This holds on the free tier exactly as it does on Pro and Business. For unpublished dissertation work, that matters.
The free tier needs no email and no account. If you are cautious about disclosing in-progress research, you can scan a draft section without us ever knowing who you are or which university you attend.
Scan history is private to your account. We do not share scan data with universities, supervisors, examination boards, Turnitin, or any third party. Your scans are not part of any institutional record, and no examiner can pull them.
Any saved scan can be deleted, and on Pro you can remove individual records. Our practices are GDPR-aware, and a standard DPA is available on Business and Enterprise for university writing centres and graduate-school cohorts.
More for university students.
The undergraduate-focused page with the four-step pre-Turnitin workflow and false-positive defence.
For college →Seven-tool ranking with Turnitin correlation and false-positive rates side by side.
See the ranking →The pre-scan workflow that catches Turnitin flags before your supervisor does.
Read the guide →Free, Starter, Pro, Business. Yearly billing saves 25%.
See pricing →Free to try. No card. 3 scans a day on the free tier.
How TextSight fits other teams and workflows.