Transparent AI detection · Privacy-first
Check If Text Was Written by AI
Analyze text in seconds and see which sentences are most likely AI-generated.
No credit card required.
Minimum 80 words. Your text is analyzed securely and never used for AI training.
Privacy-first
Anonymous scans are never stored. Your text is never used to train AI.
Secure processing
Encrypted in transit, isolated per account, keys never exposed to the browser.
Fast analysis
Results in seconds, with sentence-level scoring for every scan.
Transparent scoring
Likelihoods and confidence levels — never an absolute verdict.
Beyond a single number
See why text looks AI-generated
Most detectors give you a percentage and leave you guessing. We show the overall likelihood, a confidence level, and a sentence-by-sentence breakdown so you can judge the evidence yourself.
Sentence-level highlighting
See exactly which sentences drive the score, colour-coded by AI likelihood.
Writing-pattern metrics
Sentence variation, predictability, repetition and vocabulary diversity — explained.
PDF, DOCX & TXT uploads
Scan documents directly and download professional PDF reports (Pro).
Batch scanning & API
Scan up to 20 files at once, or integrate detection into your product (Business).
- Human likelihood
- 28%
- Confidence
- 84%
I rewrote the intro three times before it felt right. Furthermore, it is important to note that effective communication plays a crucial role in fostering collaboration. Teams that leverage the right tools can navigate change more effectively.
Illustrative example
How it works
Built on a specialized detection model
We don't ask a chatbot whether text “sounds like AI”. Every scan runs through a dedicated detector, then we add explainable writing metrics.
- 1
Paste or upload
Add at least 80 words. Longer samples generally provide a more meaningful analysis.
- 2
Detection model scores it
A classifier estimates overall AI likelihood; each sentence is scored individually.
- 3
Review with context
Highlights, confidence and writing patterns help you interpret — not just accept — the result.
Limitations you should know
- No AI detector is 100% accurate — results are estimates based on statistical signals.
- Short, heavily edited, or highly formulaic text is harder to classify.
- Non-native English writing is more prone to false positives.
AI detection systems can produce false positives. Use results as one signal rather than definitive proof, and never as the sole basis for academic, employment or disciplinary decisions.
How we handle your text
- No training. Your text is never used to train AI models — ours or anyone else's.
- Anonymous = not stored. Scans without an account are analyzed and discarded.
- Private history. Signed-in scans are visible only to you, and can be disabled or deleted.
- Never public. Scans and reports are never indexed or shared.
Pricing
Start free. Upgrade when you need more.
Simple plans with no hidden fees. Cancel anytime.
Free
For quick, occasional checks.
- 3 scans per day
- Up to 2,500 words per scan
- Overall AI likelihood
- Basic sentence highlighting
- No credit card required
Pro
For writers, editors and educators who check text every week.
or $149/year (save 38%)
- 200,000 words per month
- Detailed sentence-level analysis
- PDF, DOCX & TXT uploads
- Scan history
- Downloadable PDF reports
- Written explanation of results
- Priority processing · No ads
Business
For teams, agencies and platforms that scan at volume.
or $399/year (save 33%)
- 1,000,000 words per month
- Everything in Pro
- Batch file scanning (up to 20 files)
- REST API access & API keys
- Usage analytics
- Higher limits
Need details? Compare plans
FAQ
Questions, answered honestly
What is an AI detector?
An AI detector estimates how likely it is that a piece of text was produced by a large language model such as ChatGPT, Claude or Gemini. Centixio uses a specialized detection model trained on human and AI-written text, and highlights the individual sentences that show the strongest AI-associated patterns.
How accurate is AI detection?
Modern detectors are reliable on longer, unedited samples, but no detector is 100% accurate. Accuracy drops on short texts, heavily edited AI output, text translated between languages, and very formulaic human writing. That is why we show a likelihood and a confidence level rather than a yes/no verdict.
Can AI detectors be wrong?
Yes. Detectors can produce false positives (flagging human writing as AI) and false negatives (missing AI text). Formal, templated, or non-native-English writing is more likely to be over-flagged. Treat any result as one signal, and look at the sentence breakdown and context before drawing conclusions.
Does Centixio store my text?
Anonymous scans are analyzed and discarded — we don't keep the text. If you are signed in, your scans are saved to your private history so you can reopen reports, and you can turn history off or delete scans at any time. We never use your text to train AI models, and scans are never public.
Can I upload PDFs?
Yes. Pro and Business plans support PDF, DOCX and TXT uploads up to 10 MB. Text is extracted securely on our servers. Scanned images without a text layer and password-protected files aren't supported.
What languages are supported?
Detection is most reliable for English. Other major languages (such as Spanish, French, German and Portuguese) can be analyzed, but results should be interpreted with extra caution because accuracy is lower outside English.
Can I use Centixio for student work?
You can use it as a conversation starter, not as evidence. AI detection results should never be the sole basis for an academic integrity accusation, a grade penalty, or an employment or disciplinary decision. Combine results with drafts, version history, and a conversation with the writer.
How much text do I need?
At least 80 words. Longer samples generally provide a more meaningful analysis — a few paragraphs (300+ words) give the most stable results. Very short snippets simply don't contain enough signal to classify reliably.