Accessibility conformance

All non-text content has a text alternative that serves the equivalent purpose.

The static scan finds no image element without an alt attribute and no icon-only control without an accessible name. more

Prerecorded audio-only and video-only content has an alternative.

There is no audio or video anywhere in the product. more

Captions are provided for prerecorded audio in synchronized media.

No synchronized media exists in the product.

An alternative or audio description is provided for prerecorded video.

No prerecorded video exists in the product.

Structure and relationships conveyed by presentation are available in the markup.

Headings are now in order on every route, each screen has a main landmark … more

A correct reading sequence can be programmatically determined.

Not evaluated. This criterion requires a screen reader or other assistive technology to be run against the interface. That has not happened.

Instructions do not rely solely on shape, color, size or position.

Not evaluated. This requires an editorial read of every instruction on every screen against its visual context. The product does use color-coded status extensively, so this is a likely finding rather than a safe one.

Color is not the only visual means of conveying information.

Not evaluated, and the interface leans heavily on accent color to carry status. more

Audio that plays automatically can be paused or stopped.

Nothing in the product plays audio.

Captions are provided for live synchronized media.

No live media exists in the product.

Audio description is provided for prerecorded video.

No prerecorded video exists in the product.

Content is not restricted to a single display orientation.

Not evaluated. This criterion requires the interface to be resized, zoomed or reflowed and inspected at each step. It has not been tested at any viewport other than a desktop browser at default zoom.

The purpose of common input fields can be programmatically determined.

Not evaluated. The product collects very little personal input, so the population of fields in scope is small, but it has not been enumerated.

Text has a contrast ratio of at least 4.5 to 1, or 3 to 1 for large text.

Not evaluated. This criterion requires measured contrast ratios against the rendered pixels. The palette is dark-on-dark by design and has never been measured. The product uses a great deal of slate-500 caption text on near-black panels … more

Text can be resized to 200 percent without loss of content or function.

Not evaluated. This criterion requires the interface to be resized, zoomed or reflowed and inspected at each step. It has not been tested at any viewport other than a desktop browser at default zoom. Much of the interface is set at ten and eleven pixels, so this is a likely finding.

Text is used rather than images of text.

Not evaluated. No image of text is known to exist in the product.

Content reflows without two-dimensional scrolling at 320 CSS pixels wide.

Not evaluated. This criterion requires the interface to be resized, zoomed or reflowed and inspected at each step. It has not been tested at any viewport other than a desktop browser at default zoom. Several screens carry six-column tile grids and wide tables.

User interface components and graphical objects have a contrast ratio of at least 3 to 1.

Not evaluated. This criterion requires measured contrast ratios against the rendered pixels. The palette is dark-on-dark by design and has never been measured. Every panel in the product is separated by a one-pixel hairline in a near-background tone … more

No loss of content when text spacing is increased to the stated metrics.

Not evaluated. This criterion requires the interface to be resized, zoomed or reflowed and inspected at each step. It has not been tested at any viewport other than a desktop browser at default zoom.

Additional content shown on hover or focus is dismissible, hoverable and persistent.

Not evaluated. The product uses hover states widely and shows additional content on hover in at least the navigation and several tables.

All functionality is operable through a keyboard interface.

Not evaluated. This criterion can only be settled by operating the interface from the keyboard alone, which no one has done end to end.

Keyboard focus can be moved away from any component.

Not evaluated. This criterion can only be settled by operating the interface from the keyboard alone, which no one has done end to end.

Single-character shortcuts can be turned off, remapped or are active only on focus.

Not evaluated. The product is not believed to register single-character shortcuts, but no one has audited the event handlers to confirm it.

Time limits can be turned off, adjusted or extended.

Not evaluated. The session has an expiry and the guided demonstration advances on a timer more

Moving, blinking or auto-updating content can be paused, stopped or hidden.

Not evaluated. The guided demonstration auto-advances and several panels animate on mount.

Nothing flashes more than three times per second.

Not evaluated. No flashing content is known to exist, but the animation set has not been reviewed against the threshold.

A mechanism is available to skip repeated blocks of content.

Not evaluated, and there is no skip link in the shell. more

Pages have titles that describe topic or purpose.

Every route in the scan returns a non-empty document title. more

Focusable components receive focus in an order that preserves meaning.

The scan confirms no element carries a positive tabindex, so the document order is the focus order. more

The purpose of each link can be determined from its text or context.

The scan finds no link without an accessible name. more

Multipoint or path-based gestures have a single-pointer alternative.

Not evaluated. No multipoint gesture is believed to exist in the product.

Functions triggered by a single pointer can be aborted or undone.

Not evaluated.

The accessible name of a control contains its visible label text.

Not evaluated. Two controls in this product were given an accessible name that is more descriptive than the visible label … more

Functions operated by device motion have a component alternative.

Not evaluated. No motion-actuated function is believed to exist.

More than one way is available to locate a page.

Not evaluated. There is a persistent navigation and there are cross-links between screens … more

Headings and labels describe topic or purpose.

The scan confirms every route now begins at a level one heading and that no heading level is skipped. more

The keyboard focus indicator is visible.

Not evaluated, and the product overrides several default focus styles in its own component kit. more

The focused component is not entirely hidden by author-created content.

Not evaluated. The shell uses a sticky header and a fixed side navigation, both of which can obscure a focused element.

Functionality using dragging has a single-pointer alternative.

Not evaluated. No dragging interaction is believed to exist.

Pointer targets are at least 24 by 24 CSS pixels or are adequately spaced.

Not evaluated. The product uses compact controls throughout, several of which are visibly under twenty-four pixels tall.

The default human language of each page is programmatically determined.

The root document declares a language attribute and every route inherits it. more

Receiving focus does not initiate a change of context.

Not evaluated. This criterion can only be settled by operating the interface from the keyboard alone, which no one has done end to end.

Changing a setting does not automatically change context unless the user is advised.

Not evaluated. Several filter controls update the view immediately on change, which may or may not be a change of context under this criterion.

Help mechanisms appear in the same relative order on each page.

Not evaluated. There is no help mechanism in the product, which may make this criterion inapplicable, but that judgment has not been made formally.

Input errors are identified and described in text.

Not evaluated. The refusal messages the product returns are descriptive, but no one has checked that every one of them is announced where the error occurred.

Labels or instructions are provided when content requires user input.

The scan found two unlabeled form controls and both have been given an accessible name. more

Information previously entered is auto-populated or available to select.

Not evaluated.

The language of each passage that differs from the page language is identified.

Not evaluated, and the product now stores policy text in six languages. more

Repeated navigation appears in the same relative order.

Not evaluated. The navigation is rendered from one shared definition, which makes conformance likely, but that is an inference rather than a check.

Components with the same function are identified consistently.

Not evaluated.

Where an input error is detected and a correction is known, it is suggested.

Not evaluated.

Submissions are reversible, checked or confirmed.

Not evaluated. The decision and scenario endpoints are irreversible once signed, and the confirmation step in front of them has never been assessed against this criterion.

No cognitive function test is required unless an alternative exists.

Not evaluated. Sign-in is an email and password form with no puzzle or transcription step, so conformance is likely, but it has not been checked.

For all user interface components the name, role and value can be programmatically determined.

The scan checks that every button and link has an accessible name and that no reference to a describing element points at an identifier that does not exist. more

Status messages can be programmatically determined without receiving focus.

Not evaluated. This criterion requires a screen reader or other assistive technology to be run against the interface. That has not happened. The product raises toasts and inline refusal messages that carry no live region.