AnyFact
HomeUse casesWas This Document Made With AI?

AI DOCUMENT · FREE PLAN

Was This Document Made With AI?

Reports, theses and presentations mix text, images and metadata. The AnyFact document detector checks each part with the right detector, page by page.

Open in beta · 10 free credits a month · Files deleted after 30 days

AI Document Detector

Open in beta

What you need

Upload a PDF, DOCX or PPTX file

PDF, DOCX or PPTX, up to 20 MB and 50 pages.

✓ 10 free credits a month✓ Files deleted after 30 days
Free plan — 10 credits every month, no credit card.

At a glance

Per page

Text, embedded images and metadata are checked and summarised page by page.

AnyFact method

3 levels

Probably human, uncertain or probably AI — never a bare percentage — with the reasons and the limits of the method.

AnyFact method

10 credits

Free every month with an account, no credit card. Uploaded files are deleted after 30 days.

AnyFact free plan

Why a document needs more than a text check

A document can contain generated text, generated images, or both — and its metadata can say which software produced it and when.

The AnyFact document detector extracts the text, embedded images and metadata of PDF, DOCX and PPTX files, sends each part to the right detector and summarises the result page by page.

Scanned documents are read with OCR, so a photographed or printed page can be checked too.

Does this sound familiar?

  • A report or thesis may have been generated.
  • A presentation reuses images of unknown origin.
  • A file’s metadata does not match its claimed author or date.

Here's what we check

01

Upload the file

PDF, DOCX or PPTX, up to 20 MB and 50 pages.

02

Text, images, metadata

Each part goes to the right detector; unusual authoring software or dates are flagged.

03

A verdict in three levels

Probably human, uncertain or probably AI — never a bare percentage — with a short explanation and the limits of the method.

SourceChecksStatus
PyMuPDF / python-docx / python-pptx extractionAI Document Detector✓
Tesseract OCRAI Document Detector✓
AI text and photo detectorsAI Document Detector✓
Document metadataAI Document Detector✓

How to read the result

Probably human

The signals point to a person. It is not a guarantee: a careful generator or heavy editing can leave few traces.

Uncertain

The signals disagree or are too weak. We say so rather than guess, and list what was found.

Probably AI

The signals point to generation. It is an indication, not proof — check the context before acting.

Sample layout. Real reports name the signals that were checked and the limits of the method.

How to check a PDF for AI, step by step

Upload the original file, then read the page-by-page summary before the overall verdict. The AI document detector accepts PDF, DOCX and PPTX files.

  1. 01Use the original export where possible. A file printed to PDF or scanned loses its text layer and much of its metadata.
  2. 02Upload a PDF, DOCX or PPTX file up to 20 MB and 50 pages.
  3. 03The text is checked block by block, embedded images go through the photo detector, the metadata is read, and scanned pages are read with OCR.
  4. 04Read the result page by page, then the overall verdict: probably human, uncertain, probably AI or mixed.
  5. 05Download the PDF report to keep with your notes.

A document analysis uses 3 credits.

What does an AI PDF detector check in each file?

It splits the file into its parts and sends each part to the detector built for it.

Part of the fileHow it is checkedWhat it can reveal
TextBlock by block, with the passage-scoring text methodGenerated sections inside an otherwise human document
Embedded imagesUp to 30 images through the photo detector; icons are skippedGenerated illustrations, charts or photos
MetadataRead from the fileThe software that produced the file, authors and dates
Scanned pagesRead with OCR, for the first 20 scanned pagesThe text of printed or photographed pages

How to read a page-by-page AI paper detector result

Look at which pages are flagged and why: text and images tell different stories. A document whose text reads as human but whose cover image is generated is mixed, and that is a perfectly common outcome.

When the verdict is mixed, list the flagged pages and ask the author about those specifically: who wrote that section, which tool made that image. Specific questions get more useful answers than a general suspicion about the whole file.

PatternLikely reading
One or two pages of text flaggedA section written or rewritten with an assistant, or a formulaic part such as a summary
Only images flaggedGenerated illustrations in a human-written document
Every page flaggedThe document was probably generated as a whole
Mostly uncertainShort text per page, many tables or heavy templating

You can also read a file’s metadata yourself: in most PDF readers and office suites it sits under the document properties. Treat what you find as context. Author and date fields are easy to change, and the “creator” is often just the program that exported the file.

What an AI document detector cannot see

The analysis covers what the file contains, within set limits. Keep these in mind when you read the result:

  • Files over 50 pages or 20 MB need to be split before upload.
  • Only the first 20 scanned pages are read with OCR, and only up to 30 embedded images are checked.
  • Templates, legal boilerplate and translated passages are predictable by nature and harder to judge.
  • Metadata shows how a file was produced and saved, not who had the ideas in it.

When a file is over the limits, check the pages that matter most for your question first, such as the body of a thesis rather than its appendices, or the slides with images of unknown origin.

Some files also carry C2PA content credentials, which are explained in our guide to C2PA content credentials. For the text itself, the guide on how to tell if text is AI-generated lists checks you can run by hand. Recruiters reviewing CVs and work samples will find a fair process on our hiring and recruiting page.

Before you check

Is the result proof that a content was made by AI?▼

No. It is an indication, not proof. Detectors can be wrong in both directions, especially on compressed, cropped, edited or very short content. Use the result to decide what to check next, not as a final judgement on anyone.

Why no percentage?▼

A single percentage suggests a precision no detector has. You get one of three levels — probably human, uncertain or probably AI — with the signals behind it and the limits of the method.

How does the analysis work?▼

When the content carries provenance data, such as C2PA content credentials or metadata, we read it first. Then an open-source detector scores the content. When the score is borderline, Claude reviews it as a second opinion, and the result is explained in plain words.

What happens to my files?▼

They are processed to run your analysis — including, for borderline cases, a review by Claude, Anthropic’s model — and deleted after 30 days.

What does it cost?▼

There is a free plan: a free account gets 10 credits every month, with no credit card. Pro ($15 a month) gives 1,000 credits a month; Business ($49 a month) gives unlimited Business Checks and AI checks under a fair-use limit.

Can I check a document longer than 50 pages?▼

Not in one analysis. Split the file into parts of 50 pages or fewer, for example by chapter, and check each part separately.

Does it check the images inside a PowerPoint?▼

Yes. Up to 30 embedded images per file go through the photo detector, and small icons are skipped.

What does document metadata tell me about AI use?▼

Metadata records things like the software that produced the file, its authors and its dates. When it names a known AI generator, that is a strong signal. More often it only adds context, and it can be edited by anyone with the file.

Can it check a scanned PDF?▼

Yes. Scanned pages are read with OCR, for the first 20 scanned pages of the file. Poor scans give the text detector less reliable text to work with.

Check a document, page by page

What you need

Upload a PDF, DOCX or PPTX file

Free plan — 10 credits every month, no credit card.

Each result is an indication, not proof.

Files deleted after 30 days.