Public documentation

Screen Reader Review aid

What the screen-reader review support workflow can organise and what it cannot prove.

Review support, not real AT validation

The Screen Reader Review aid is a workflow for organising review evidence. It is not a replacement for testing with real screen readers and does not prove NVDA, JAWS, VoiceOver, TalkBack, Narrator, switch control, magnification, keyboard-only, or user-task behaviour.

What it supports

  • Preparing a reviewer for real assistive technology testing.
  • Organising expected reading order, visible content checks, notes, and evidence.
  • Reviewing likely issues around headings, landmarks, links, forms, names, roles, and interaction flow.
  • Exporting structured review notes where supported.

Inputs and outputs

Input or outputPurposeBoundary
Rendered page structureSupports expected reading-order planning.Does not equal a real accessibility tree from every AT/browser pair.
Visible content checklistHelps compare expected content with what a reviewer should hear or reach.Needs real AT and keyboard confirmation.
Virtual review outputSupports repeatable review preparation.Not proof of real AT behaviour.
Reviewer notesCaptures human observations and next steps.Can include sensitive page context.
Exported evidenceSupports handoff to QA, engineering, or accessibility specialists.Share only with authorised recipients.

Responsible workflow

  1. 1

    Run the page scan

    Collect current-page scanner evidence and diagnostics.

  2. 2

    Open Screen Reader Review

    Use the aid to identify likely reading-order and interaction paths.

  3. 3

    Test with real assistive technology

    Run the target AT/browser combination for the user journey.

  4. 4

    Record differences

    Capture where real AT output differs from the aid or where manual testing finds additional issues.

  5. 5

    Export evidence carefully

    Share only with authorised recipients because exports can include page context and notes.