Detect if a PDF is scanned

Analyze a PDF to estimate if it comes from a scan (images on pages, absence of fonts, CCITT/JBIG2 indicators) and export the diagnosis in JSON, locally in your browser.

Import a PDF

No data is sent to a server: everything is done in your browser.

← Back to PDF tools

Use cases

Decide on OCRQuickly know if OCR is needed before indexing/full-text search.
Accessibility CheckIdentify non-selectable "image" PDFs (often inaccessible) to correct.
Extraction PipelineAutomatically route documents: direct extraction if text is present, OCR otherwise.
Scanning QAIdentify low-quality pages (too little text, dominant images) in a batch of scans.

What you get

A JSON diagnostic per page: presence of text, proportion of image content, "scan" indicators (based on heuristics).

Indicators to prioritize OCR (pages to process, already exploitable pages).

Limits and points of attention

Heuristic detection: some graphic PDFs (catalogs, slides) may look like scans without being so.
Some PDFs may contain "hidden text" (OCR already done): the tool detects it but quality may vary.
The actual quality of the OCR will depend on resolution, contrast, and language — not finely evaluated here.

Frequently asked questions

Can a scan contain selectable text?

Yes, if OCR has been applied: the image remains, but a text layer is added.

Does the tool perform OCR?

No. It detects and diagnoses. OCR is a separate step.

Why is a page classified as "scan" when I see text?

The visible text may be an image (rendered), not vector text. The diagnosis is based on the internal content.

Tip

Before investing extraction time, detect scans: it’s the simplest way to avoid "hunting" for text that doesn’t exist without OCR.