Accessibility statement

1. The standard

We aim to meet the Web Content Accessibility Guidelines (WCAG) 2.1 at level AA, in the website, the installed app and the iPhone, iPad and Android apps, which show the same pages. Both the light and the dark theme are held to it, and so are the phone and the tablet layouts.

2. What helps you read

• The reader lets you change the text size, the typeface (including one designed for readers with dyslexia), the line width and the line spacing, and the page reflows. Pages zoom with the browser, and nothing depends on a particular screen width.
• Light and dark themes, following the device unless you pick one. The PDF reader can show pages dark.
• Read aloud (where it is switched on for you) turns a web page, book or PDF into speech with an AI voice, and follows along on the page. Voice notes are transcribed, so a comment is also text.
• Pages, menus and dialogs can be used with a keyboard. Dialogs keep the focus inside while they are open and give it back when they close, and Escape closes them. The reader has shortcuts for the common actions. Controls have names a screen reader can announce.

3. How we check

• Automated checks on every change. The test suite runs the axe accessibility engine over the main screens, in the light and the dark theme: the home page and sign-in, the list of groups, a group, a meeting, the reader with a web page and with a PDF (with a discussion open), the comments, the read aloud player, dialogs, settings, people, and the legal pages including this one. A serious or critical finding fails the build, so it can't be deployed. The suite runs on Chromium and on WebKit (the engine of iPhones and iPads) and on tablet and phone sizes.
• Manual checks with screen readers are still to come. Automated tools find only part of the problems, so we will go through the main tasks with VoiceOver on iPhone and Mac and with TalkBack on Android, and update this page with what we learn. Until then, assume problems we haven't found exist.

4. Known limitations

• PDFs without a text layer. A scanned PDF is a picture. Without text in it, a screen reader has nothing to read, you can't select a passage to comment on, and read aloud has nothing to say. Diplet doesn't read pictures of text (no OCR). Add a text version of the reading (paste it, or use a web page or book) in the meantime.
• What is in a reading is up to who added it. Diplet shows the pages, PDFs and books your group adds as they are, apart from removing what could run code. Images in them have the descriptions the author gave them, if any, and a PDF has the structure its maker gave it.
• Highlights in the text. The passages that have comments are marked on the page with colour. A screen reader doesn't announce the marks; the comments panel lists every discussion with the passage it is about, in order, and that is the accessible way to find them.
• Selecting a passage. Commenting on a passage starts from a selection of text, which is made with the browser's own selection (a long press on a phone, the mouse on a computer). With only a keyboard that needs a browser's caret browsing (F7 in Chrome and Firefox), and on a PDF page the selection follows the text layer, which can differ from how the page looks. A comment on the whole reading needs no selection, and works with the keyboard alone.
• The AI voice and AI answers are generated and can be wrong or oddly pronounced, especially with names, formulas and other languages.

5. Tell us about a problem

If you can't use something, or it is harder than it should be, please write to our contact address. Say what you were doing, on which device, with which browser or app and, if you use one, which assistive technology. We will answer, and say what we can do and by when. If you have a reading we can't make accessible, we will help find another way to get you the text. Barriers that stop you from taking part in a group are fixed first.

See also the Terms of Service and the Privacy Policy.