Frequently Asked Questions
Everything you need to know about checking your PDF documents for AI-generated content, interpreting accuracy scores, and maintaining data privacy.
General Questions
It’s incredibly simple. Go to our homepage, and drag and drop your PDF file into the upload zone (or click to browse). Our tool will automatically extract the text from the document, run our AI detection algorithms, and provide you with a probability score and sentence-level highlights within 10-15 seconds.
Unlike other tools, you do not need to manually copy and paste text from your PDF.
Yes. We offer a robust free tier that allows anyone to upload a PDF (up to 10MB) and receive an overall AI probability score along with a preview of the sentence-level analysis. You do not need to create an account or provide a credit card to use the free scanner.
We also offer a premium tier for users who need full, unredacted sentence-by-sentence reports and downloadable PDF certificates.
We natively support .PDF, .DOCX (Microsoft Word), and .TXT files. If you have an unsupported file format (like an RTF or Apple Pages file), you can either export it as a PDF or use our “Paste Text” tab on the homepage.
Detection & Accuracy
Our detection engine is trained to identify linguistic patterns from all major Large Language Models (LLMs). This includes:
- OpenAI: ChatGPT, GPT-3.5, GPT-4, GPT-4o
- Anthropic: Claude 2, Claude 3 (Opus, Sonnet, Haiku), Claude 3.5
- Google: Gemini 1.0, Gemini 1.5 Pro and Flash
- Open Source: Meta LLaMA, Mistral, DeepSeek
On standard, unedited AI-generated text, our accuracy rate is above 95%. However, accuracy depends heavily on the length of the document and the level of human editing.
The detector works best on documents with 500+ words. If AI-generated text is heavily edited by a human or run through advanced paraphrasing tools, the detection confidence may drop. We recommend using our tool as a strong indicator rather than absolute proof.
This is known as a “false positive.” AI detectors look for low perplexity (predictable word choices) and low burstiness (uniform sentence structures).
Highly formal academic writing, technical reports, and heavily structured essays often exhibit these same traits naturally. If you write with very rigid, repetitive sentence structures, the algorithm may misinterpret it as AI. We always advise educators to look at the context and use the tool’s results as a conversation starter with the student.
Yes. If you upload a flattened PDF (like a scanned document where the text cannot be highlighted), our system automatically runs an Optical Character Recognition (OCR) process to extract the text from the image layers before running the AI analysis.
Privacy & Security
Absolutely. We operate a strict Zero Retention policy. Your documents are processed entirely in memory (RAM) and are permanently deleted the exact millisecond the detection scan is finished. We never save your files to a hard drive or database.
No, never. Unlike some free services that monetize by scraping user data, we do not use your documents to train our detection algorithms, nor do we sell or share your text with third-party AI companies like OpenAI or Google.
Yes. Because we do not require account creation for basic scans and because we instantly delete all uploaded data without storing it, we are natively compliant with the General Data Protection Regulation (GDPR). For more details, see our Privacy Policy.
Still have a question?
If you couldn’t find the answer you were looking for, our team is here to help.
Contact Support