Statistical and stylistic analysis
It spots anomalies typical of “non-human” texts.
Comparison
Quick Analysis or Deep Analysis?
Both analyses use the same AI to detect Claude, Gemini & ChatGPT text in German and English – the Deep Analysis additionally explains the result through detailed style checks and data forensics.
Performance comparison
The AI text detector models by PlagiatGPT
Quick Analysis and Deep Analysis run on the same AI detector – so the detection rate is identical. What matters is choosing the right model:
- General for website copy or blog posts
- Education for texts from an educational context such as school or seminar papers
Detection rate for AI-generated texts
higher is better · at under 1% false-positive rateEN / General
Blogs, articles, professional & everyday texts
Detection rate
94.4 %
DE / General
Blogs, articles, professional & everyday texts
Detection rate
87.8 %
EN / Academic
Research papers, essays, academic submissions
Detection rate
89.9 %
DE / Academic
Term papers, theses, academic essays
Detection rate
90.1 %
False-positive rate
lower is better
Detection rate
<1 %
The detection rates come from the out-of-distribution benchmark of our AI detector models (V3.1, July 2026), measured on unseen, external datasets at under 1% false-positive rate. Depending on the text category, the detection rate is between 87.8% and 94.4% (German/general 87.8%, English/academic 89.9%, German/academic 90.1%, English/general 94.4%). Quick Analysis and Deep Analysis use the same detector, so the figures apply to both. Note: the success of AI detectors depends heavily on the category of the text. Deliberate obfuscation and alterations to AI-generated texts can reduce the reliability of the AI classification. Highly technical texts as well as purely mathematical, abstract and unstructured texts are excluded from AI analysis.
Reliable results
Why the Deep Analysis?
We believe that the best AI detector is, in the end, judged by a human. That’s why the Deep Analysis additionally shows you:
Data forensics
It detects machine-generated documents such as Word or PDF.
Sentence-by-sentence analysis
Have suspicious sentences highlighted.
In practice, that means a tool with which teachers or editorial teams can form a well-founded judgement using their own experience.
Direct comparison
Quick Analysis vs. Deep Analysis
Quick Analysis
Free · no sign-up
Deep Analysis
From €13 / month
PriceFreeFrom €13 / month
AI detectorPlagiatGPT Frontier AI text detector V3.1PlagiatGPT Frontier AI text detector V3.1
tested AI detection rate87.8–94.4%87.8–94.4%
False-positive rate<1%<1%
ResultScore 1-10 (without an account), exact percentage with a free accountExact percentage + sentence-level highlighting + rationale
Text lengthShort excerptWhole documents
QueueBehind paid plansPriority
AI detector model for academic texts (papers, semester assignments, homework) With a free account Yes
Stylometry analysis – writing-style fingerprint No Yes
Highlighting of suspicious passages No Yes
Store test results encrypted No Yes
Compare multiple documents in folders No Yes
File forensics (PDF/DOCX) NoPro & Max
Upload up to 40 papers simultaneously and instantly see anomalies NoPro & Max
Built for professional use
Who is PlagiatGPT for?
Teachers & lecturers
Detect AI-generated text in a GDPR-compliant way and gain a solid basis for conversation.
Why teachers rely on PlagiatGPT.Education providers & universities
Privacy-compliant and ready to use right away – hosted in Germany, with legal documents for your data protection officer for all 16 federal states, simply available to download.
How education providers and schools benefit from PlagiatGPTSEO & content teams
Search engines penalise AI-generated content. Dominate the rankings and avoid risks by checking your content.
Why SEO teams rely on PlagiatGPTStudents
Check before submitting how AI detectors rate your work – and walk into the assessment with no nasty surprises.
Check your paper for free now.File forensics
Metadata doesn’t lie.
Every document leaves invisible traces – Word files from ChatGPT, Gemini and the like give themselves away in the metadata. Our file forensics (Pro and Max plans) reads these traces and shows whether a document was really typed or machine-generated – regardless of how good the text sounds.
File forensics
Checking metadata and hidden signals in the document.
Automatically created Word document
This Word document does not appear to have been created by a human. The file was either AI-generated, converted, exported or otherwise produced digitally.
Frequently asked questions about the Deep Analysis
No – a detection result is not legal proof, but a well-founded indicator. The Deep Analysis provides rationale as a basis for conversation. Final consequences – for example in examination or disciplinary proceedings – are always decided by humans based on all available circumstances.
The Deep Analysis identifies conspicuous passages and provides clues to unusual indicators. Scores between 40–69% are considered indeterminate and should not be treated as a verdict. If there is good reason to suspect a false alarm, a direct conversation with the person is the best next step.
German and English – each with separately trained models for general and academic texts. This gives four calibrations:
DE/General, DE/Academic,
EN/General, EN/Academic.
The language is detected automatically - you do not have to set anything.
Yes, to a degree. The detection models were specifically trained on paraphrased texts and texts processed by humanizer tools. Structural and stylometric patterns remain detectable even after rephrasing; with heavily manually edited texts the rate drops.
The Deep Analysis processes PDF, DOCX, TXT and Markdown. For Word and PDF files, file forensics additionally kicks in (Pro & Max), checking metadata for signs of manipulation.
Quick Analysis: no – discarded immediately, fully anonymous. Deep Analysis: AES-256-encrypted, visible only to you, never used for training. All on servers in Germany.
Ready for the Deep Analysis?
Sentence-level highlighting, file forensics and one of the best detection rates on the market.