Catalog and metadata
File-wide conditions
PDF accessibility checker API with traceable evidence
The PDF accessibility checker API accepts an uploaded PDF and the technical profiles your workflow needs. It returns machine-verifiable PDF/UA results alongside structure evidence, heuristics, and explicit human-review work.
For CMS, document-generation, and publishing systems that need a PDF preflight gate
Illustrative PDF finding
IllustrativeThe PDF has no usable tag structure
blocker · deterministic- Page
- Document-wide
- Locator
- PDF catalog / StructTreeRoot
- Observed
- missing
- Validator
- PDF/UA profile + custom structural rule
- Profiles
- PDF/UA-1 · PDF/UA-2 · document mappings
- Locator
- Page · object · structure · MCID when available
- Boundary
- Machine validation is not complete accessibility
Inspect the PDF as structure, objects, pages, and human reading experience.
The evidence model keeps machine-verifiable document conditions close to the parts that still require interpretation.
Technical centerPDF preflight
Tags and relationships
Machine-observable semantics
Content and annotations
Page and object locators
Meaning and order
Deliberate evaluation
Keep the PDF repair location with the rule.
Page numbers and object references make a structural failure more actionable than a document-level pass/fail label.
Figure alternative description is missing
detected failure- Page
- 18
- Object
- Figure tag · MCID 27
- Observed
- Alt entry absent
- Automation
- Deterministic presence check
Repair pathAdd a purposeful alternative in the source when possible, regenerate the tagged PDF, then have a person verify that the description conveys the figure’s meaning.
Choose the technical profile behind this capability.
Select PDF/UA, WCAG-oriented document checks, or another supported profile according to the project's technical target. Each keeps a separate version and coverage record.
Separate PDF/UA machine validation from document usability.
A validator can establish technical conditions in the file. It cannot decide whether every semantic choice works for a reader.
Machine-verifiable
- PDF/UA declaration and validator rules
- Presence and shape of the tag tree
- Selected metadata and structural conditions
- Object-level evidence where extractable
Human-verifiable
- Meaningful alternative descriptions
- Logical reading order in context
- Complex table usability
- Complete screen-reader experience
Put traceable PDF evidence before the publish action.
Scan a PDF in a free project, review its object-level findings, and connect the API when repaired files need a consistent retest.
Check a website freeCreate an account
Automated preflight is not an accessibility certification or legal determination.