Voice (dictate, read aloud) and Draw a picture over /v1, with the app's models scope

The server kept voice and images on /v1 (models scope, which the app's token
already asks for); /me's services and scopes decide what is offered. Rust
sends the recording to /v1/audio/transcriptions, reads /v1/audio/speech back
as a data: URL (CSP media-src data:), and for /v1/images/generations takes
only the id of each of the person's own files from the answer and fetches
/v1/files/<id>/content from the app's own server. The microphone is allowed
only within seconds of the page's own Dictate (voice_arm); WebKitGTK's
media stream is switched on through its settings getter.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

ui/js/chat.js

@@ -22,6 +22,7 @@ import { unarchive } from "./archived.js"
 import { isProgress, progressLine, progressTicks } from "./progress.js"
 import { aimTerminal, terminalButton, terminalOpen, toggleTerminal } from "./terminal.js"
 import { scopePopover } from "./scope.js"
+import { readAloudButton } from "./voice.js"
 
 const MODES = ["manual", "edit", "auto", "plan"]
 
@@ -292,6 +293,7 @@ function assistantMessage(m, chat) {
         "div",
         { class: "msg__actions" },
         action(ICONS.copy, t("Copy"), () => copyText(replyText(m))),
+        readAloudButton(m.id, () => replyText(m)),
         // Only the last reply: CSS shows it there and nowhere else (app.css, .msg__regen).
         action(ICONS.refresh, t("Regenerate reply"), () => regenerate(m), "msg__regen"),
       ),