Controlled PDF fixtures

Three document types. Three different evidence boundaries.

These fictional files let BarrierTrace develop and demonstrate PDF inventory and triage without implying that automated inspection proves PDF or WCAG conformance.

Native text detected

Native-text service guide

Contains extractable text, a document title and a language setting. It deliberately has no tagged structure, so semantics and reading order still need manual assessment.

Prototype result: Text extraction works. OCR is not required.

Open test PDF

Text and image detected

Mixed-content service report

Contains ordinary text plus a raster chart. A parser can identify both content types but cannot derive a trustworthy alternative description from structure alone.

Prototype result: Text extraction works. The image needs vision or manual interpretation.

Open test PDF

Image-only page detected

Image-only scanned notice

The entire page is a single image with no extractable text layer. This represents scanned correspondence and legacy image-only documents.

Prototype result: OCR or vision is required, followed by human verification.

Open test PDF