SoulkeyResearch

Free PDF text-layer check

Text, scan,
or something in between?

Check which pages contain extractable text before choosing a PDF-to-CSV workflow. Your selected PDF is processed in this browser and is not uploaded.

Local processing · No account needed

Choose a PDF on your device

Up to 15 MiB · 50 pages · 20-second analysis limit

We do not upload, save or display your document’s text. Only page-level character counts are shown. Clear the result when you’re finished.

Loading the checker…

Or try one of our sample PDFs

Invented data only. These buttons download a sample from this site and check it locally.

Choose your next step

Start with a free workflow.

No text or mixed pages

Check whether OCR is needed

Understand image-only pages and the limits of a text extractor. Test a small sample and inspect the result.

Text present, table broken

Keep headers out of your data

Explore free pdfplumber and pdftotext workflows, with a specific rule for repeated column labels.

After any conversion

Check IDs, not just totals

A correct sum can hide incorrect records. Use our field, structure and identity checklist.

Method and limits

A text signal, not a verdict.

The checker uses Mozilla PDF.js 6.3.289, bundled on this site. It reads text items sequentially in an isolated web worker and counts non-whitespace characters. It does not retain extracted strings in the result, render pages, execute PDF scripts, perform OCR, or identify tables.

A page with 0 characters is labeled “No text found.” A page with 1–39 characters is flagged for review because a page number or heading can coexist with scanned content. At 40 or more characters, we report that a text layer was detected. None of these categories guarantees complete or correct text.

Password-required PDFs, unreadable files and files over the size/page limits are not checked. Processing is stopped at the time or text-complexity limits, and incomplete checks do not receive a success verdict. Compressed PDFs can still use substantial browser memory before a limit is reached. Use a current browser; browser scheduling and background-tab throttling can affect the time cap.

The parser’s resources are self-hosted. No third-party CDN, analytics or document-processing API is used. Loading this page and public demos still makes ordinary requests to this site. Hosting security features and external links have their own practices; see privacy and document handling.

Diagnostic v1.0 · 30 September 2026 · PDF.js Apache 2.0 license · Recorded diagnostic test results