feat: voice dictation (STT wave 1 — Deepgram cloud)

Two-band architecture from docs/research/stt.md: renderer captures mic
audio in an AudioWorklet (linear16 16kHz mono, ~128ms frames, per-quantum
RMS for the meter) and streams it directly to Deepgram over WebSocket —
audio never loops through our server. The real key stays server-side;
/api/stt/token mints a short-lived grant token (TTL 300s) per run.

Patterns ported from the owner's pi dictate extension (tuned in real
use): generation counter guarding every async handler, the
latest-RMS-never-resets meter (-50..-10dB window, 60ms tick), mic-first
stop + CloseStream + 3s safety net, endpointing=300 with the
utterance_end_ms=1000 backstop.

UI (owner placement): mic button in the stats bar's middle slot on all
three surfaces (home graph-view, chat, draft); while recording the meter
unfurls to fill the row's free width with the interim ghost at its
flank. Esc ends dictation from the composer. Token mint comes before
getUserMedia so a keyless app fails fast without a TCC prompt; one
automatic reconnect on abnormal socket close, errors surface inline.

Live-smoked the full error path in the dev shell (bar renders on all
three pages, 403 grant failure shows inline, mic untouched). The full
audio path needs a Deepgram key with Member permission to mint tokens.

components/stats-bar.tsx

@@ -33,6 +33,7 @@ export function StatsBar({
 	settings,
 	onSettingsSaved,
 	leftSlot,
+	dictationSlot,
 }: {
 	turns: TurnUsage[];
 	chatId: string;
@@ -41,6 +42,11 @@ export function StatsBar({
 	/** Pre-chat only: replaces the (empty) session-data summary — home and
 	 *  the /chat/new draft mount the knowledge-context slider here. */
 	leftSlot?: ReactNode;
+	/** The dictation strip (STT wave, 2026-09-17): the row's middle is dead
+	 *  space between the session data and the model control — the mic
+	 *  button lives there, and while recording the meter unfurls to fill
+	 *  whatever width the row has (owner placement). */
+	dictationSlot?: ReactNode;
 }) {
 	const [open, setOpen] = useState(false);
 	// Same dropdown discipline as the model control (owner, df4433f treatment,
@@ -155,6 +161,12 @@ export function StatsBar({
 				)}
 			</div>
 			)}
+			{/* The middle slot: dictation (owner, 2026-09-17). flex-1 lets the
+			    meter expand into the row's free width; justify-center keeps the
+			    lone idle mic button visually centered. */}
+			{dictationSlot && (
+				<div className="flex min-w-0 flex-1 items-center justify-center">{dictationSlot}</div>
+			)}
 			{/* The model + thinking control — one controller for the conversation
 			    agents (prober/teacher); the record's `turns` above still folds
 			    the ACTUAL models billed, regardless of what the picker says. */}