aropjoe · Oct 10, 2026 · Oct 9, 2026 · Oct 10, 2026
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -7,7 +7,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [1.0.0-alpha.3] - 2026-10-10
+
 ### Added
+- Accessibility: Duct aims to meet WCAG 2.2 AA, and publishes a conformance report (VPAT®) at duct.tensflare.com/legal/accessibility. Text and field borders meet the contrast minimums; focus is always visible; there's a Skip to content link; the documents list is a table for screen readers; the number of results is announced. In a document, ↑ and ↓ select a paragraph and Enter adds it to a notebook, and ⌘⇧N / Ctrl+Shift+N adds any selection. Everything reflows at 400% zoom. The / shortcut can be turned off (Settings › Features › Keyboard), and the mascot stops moving within 5 seconds. `npm run a11y` checks every main screen with axe-core, and CI runs it.
 - Search results are grouped by document. A PDF that matches on several pages is one result with its best passage, links to the other pages ("Also on p. 2, p. 7") and its other passages a click away; the heading counts documents and passages. Quick search shows one row per document.
 - Open shows the document inside Duct, on its own, with your notebooks beside it; **Open in app** hands it to the program it belongs to. Closing one of two documents in the workspace leaves the other full width, and **One / Side by side / Stacked** switches between them.
 - Selected text can go to any notebook: the pop-up says **Add to "<notebook>"**, and its menu lists every notebook you can add to, or starts a new one named after the document.

diff --git a/assets/ui/app.js b/assets/ui/app.js
@@ -732,6 +732,8 @@ function renderResults() {
   const docs = state.cards.length
   $('#resultsTitle').textContent = !results.length ? 'No results for “' + state.query + '”'
     : plural(docs, 'document') + (results.length > docs ? ' · ' + plural(results.length, 'passage') : '') + ' for “' + state.query + '”'
+  // Screen readers hear the count when a search finishes (WCAG 4.1.3).
+  $('#resultsStatus').textContent = $('#resultsTitle').textContent
   $('#results').innerHTML = state.cards.map((idx, card) => {
     const r = results[idx[0]]
     const c = r.chunk

diff --git a/assets/ui/index.html b/assets/ui/index.html
@@ -133,6 +133,7 @@ <h1 id="homeTitle"></h1>
         <section id="viewResults" class="view" hidden aria-label="Results">
           <div class="results-head">
             <h2 id="resultsTitle"></h2>
+            <p id="resultsStatus" class="sr-only" role="status" aria-live="polite"></p>
             <div class="scope-chips" id="scopeChips"></div>
             <div class="results-tools">
               <button class="btn btn-sm" data-action="show-collected" id="collectedBtn" hidden>Collected <span class="count" data-bind="collectedCount">0</span></button>

diff --git a/docs/accessibility.md b/docs/accessibility.md
@@ -0,0 +1,33 @@
+# Accessibility
+
+Duct aims to meet WCAG 2.2 at level AA. The conformance report (VPAT® 2.5, WCAG edition) is published at
+https://duct.tensflare.com/legal/accessibility/ and must stay true: when a change affects a criterion, update it
+(`duct-site/src/accessibility.mjs`).
+
+## Automated
+
+`npm run build && npm run a11y` (`scripts/a11y-electron.cjs`) starts Duct with a small sample library, opens every main
+screen in Chromium (Electron) and runs axe-core with the WCAG 2.2 A and AA rules. Any violation fails it, and CI runs it
+on macOS for every pull request. Add a screen to the list in that script when you add one to the app.
+
+## By hand, before a release
+
+axe-core can't judge these, so check them when the related screens change:
+
+- **Keyboard only:** search, open a result, step through matches (Enter, Shift+Enter), switch layouts, close a pane,
+  and add a quote: Tab to the document, ↑ ↓ to select a paragraph, Enter. ⌘⇧N / Ctrl+Shift+N adds any selection,
+  ⌘⇧M / Ctrl+Shift+M opens the notebook menu.
+- **Focus** is always visible, and dialogs give focus back when they close.
+- **Reflow:** at 400% zoom in a 1280-pixel window (320 CSS pixels), nothing scrolls sideways except wide tables.
+- **Text spacing:** with line height 1.5, letter spacing 0.12em, word spacing 0.16em and paragraph spacing 2em,
+  nothing is clipped.
+- **Contrast:** new colours meet 4.5:1 for text and 3:1 for field borders on every panel they sit on
+  (`--muted` and `--field-border` in `assets/ui/app.css` are the floors).
+- **Motion:** animations stop within 5 seconds and don't play with reduced motion.
+- **Status messages** use a live region (`role="status"`).
+
+## Known gaps
+
+- Selecting part of a paragraph needs a pointer or a screen reader's own selection (WCAG 2.1.1, partially supported).
+- Resizing two documents to an exact size without dragging needs the arrow keys (2.5.7, partially supported).
+- Not yet tested with people who use screen readers every day; planned before 1.0.
diff --git a/package-lock.json b/package-lock.json
diff --git a/package.json b/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@docfide/duct",
-  "version": "1.0.0-alpha.2",
+  "version": "1.0.0-alpha.3",
   "description": "Document intelligence pipeline — extract, chunk, embed, and search any document format",
   "type": "module",
   "bin": {