Deck: echo what voice awareness hears into the feed

README.md

@@ -87,7 +87,9 @@ Quit Hype Ghost.
   OBS plugin (local whisper.cpp — audio never leaves your PC), add its Transcription filter
   to your mic source, output to a text file or a (hidden) text source, and point the wizard's
   voice-awareness step (or the `transcript` config section) at it. Then answering the ghost
-  out loud *is* replying.
+  out loud *is* replying. Everything it hears also shows up as faint 🎙 lines in the deck
+  feed (toggle in Settings → Voice), so when the cast reacts oddly you can see exactly what
+  the transcription thought you said.
 - **Party / co-op audio (second channel):** streaming alongside friends? Route their audio
   (Discord/TeamSpeak/party chat) to a separate OBS source, add a *second* LocalVocal
   Transcription filter to it, output to a **different** file/text source than your mic, and
@@ -124,6 +126,7 @@ Quit Hype Ghost.
 | `twitch.decapi` | Fetch viewer count + stream info keylessly via [DecAPI](https://decapi.me), a community-run Twitch API proxy (default true). Only your public channel name is sent to it. Set false to opt out. |
 | `twitch.clientId` / `clientSecret` | *(Optional, power users)* App credentials from dev.twitch.tv — when set, the official Helix API is used directly instead of DecAPI. |
 | `transcript.mode` | `off`, `file` (tail LocalVocal's .txt/.srt output), or `textSource` (poll a text source over OBS WebSocket). |
+| `transcript.showInFeed` | Echo what voice awareness hears into the deck feed as faint 🎙 lines (party audio as 🎧) so you can spot mishears the cast might be reacting to (default true). Deck only — never in the cast's chat history, the recap, or the on-stream overlay. |
 | `transcript2.*` | *(Optional)* A **second** transcription channel for party/co-op audio (a separate audio device with its own LocalVocal filter). Same `mode`/`file`/`textSource`/`pollSeconds` as `transcript`, plus `label` — how the cast refers to those people (e.g. "my co-op squad"). Treated as *other people*, never as the streamer. |
 | `cadence.soloSeconds` / `quietSeconds` | Average gap between messages when alone / when real viewers are present. |
 | `cadence.jitter`, `burstChance`, `lullChance` | Natural rhythm: ±jitter on normal gaps, occasional quick bursts (0.3–0.6x) and long lulls (1.6–3x). |

config.example.json

@@ -94,7 +94,8 @@
     "mode": "off",
     "file": "",
     "textSource": "LocalVocal Captions",
-    "pollSeconds": 2
+    "pollSeconds": 2,
+    "showInFeed": true
   },
   "transcript2": {
     "mode": "off",

public/dashboard.html

@@ -120,6 +120,14 @@
   .msg.streamer .meta { flex-direction: row-reverse; }
   .msg.streamer .nm { color: var(--accent-2); }
   .msg.streamer .bubble { background: var(--accent-soft); border-color: var(--accent-line); border-radius: var(--r-lg); border-top-right-radius: 4px; }
+  /* "heard" rows — what voice awareness transcribed. Faint and italic on
+     purpose: this is sensor readout (possibly misheard), not chat. Mic lines
+     sit right like your typed messages; party lines sit left (other people). */
+  .heard { align-self: flex-end; display: flex; align-items: baseline; gap: 8px; max-width: 640px; font-size: var(--fs-sm); font-style: italic; color: var(--faint); animation: rise .28s ease-out; }
+  .heard.party { align-self: flex-start; }
+  .heard .ic { font-style: normal; flex: none; }
+  .heard .txt { overflow-wrap: anywhere; min-width: 0; }
+  .heard .tm { font-style: normal; font-size: var(--fs-xs); flex: none; }
   .moment-card { align-self: center; display: flex; align-items: center; gap: 10px; padding: 7px 15px; border-radius: var(--r-full); background: var(--glass); border: 1px solid var(--accent-line); font-size: var(--fs-sm); animation: pop .4s cubic-bezier(.2,1.3,.5,1); }
   @keyframes pop { from { opacity: 0; transform: scale(.85); } to { opacity: 1; transform: none; } }
   .moment-card .spark { background: var(--aurora); -webkit-background-clip: text; background-clip: text; color: transparent; font-weight: 800; }
@@ -260,7 +268,8 @@
       next:'up next', obsOk:'OBS connected', obsNo:'OBS offline', brainOk:'Brain ready', brainNo:'No brain', viewersManual:'manual',
       micOff:'off', micQuiet:'quiet', micWait:'waiting', party:'Party', partyQuiet:'quiet', partyWait:'waiting', chatQuiet:'quiet', chatActive:'active', lost:'Connection lost — reconnecting…',
       msgs:'msgs', watching:'watching', idle:'idle', noMoments:'No moments yet — the cast flags clip-worthy plays here.',
-      previewBody:'Preview mode — your cast has no brain yet, so the room stays quiet.', previewCta:'Connect one in Settings' },
+      previewBody:'Preview mode — your cast has no brain yet, so the room stays quiet.', previewCta:'Connect one in Settings',
+      heardTip:'What voice awareness transcribed — the cast may be reacting to this. If it misheard you, that explains their reaction. (Deck only — never shown on stream.)' },
     es: { brain:'Cerebro', viewers:'Espectadores', mic:'Micro', chat:'Chat', cast:'El elenco', energy:'Energía',
       chill:'Tranqui', balanced:'Normal', hype:'Máx', room:'Sala', modeAuto:'Auto', modeSolo:'Solo', modeViewers:'Público',
       say:'Di algo', pause:'Pausar', resume:'Reanudar', moments:'Momentos', recap:'Exportar', send:'Enviar', you:'Tú',
@@ -269,7 +278,8 @@
       next:'siguiente', obsOk:'OBS conectado', obsNo:'OBS desconectado', brainOk:'Cerebro listo', brainNo:'Sin cerebro', viewersManual:'manual',
       micOff:'apagado', micQuiet:'silencio', micWait:'esperando', party:'Grupo', partyQuiet:'silencio', partyWait:'esperando', chatQuiet:'tranquilo', chatActive:'activo', lost:'Conexión perdida — reconectando…',
       msgs:'msjs', watching:'mirando', idle:'inactivo', noMoments:'Aún no hay momentos — el elenco marcará jugadas destacadas aquí.',
-      previewBody:'Modo vista previa — tu elenco aún no tiene cerebro, así que la sala sigue en silencio.', previewCta:'Conéctalo en Ajustes' },
+      previewBody:'Modo vista previa — tu elenco aún no tiene cerebro, así que la sala sigue en silencio.', previewCta:'Conéctalo en Ajustes',
+      heardTip:'Lo que la escucha de voz transcribió — el elenco puede estar reaccionando a esto. Si te entendió mal, eso explica su reacción. (Solo en la consola — nunca en el stream.)' },
   };
   let L = STR.en;
   const T = (k) => (L[k] ?? STR.en[k] ?? k);
@@ -393,6 +403,24 @@
     renderMoments();
   }
 
+  // What voice awareness heard (mic or party channel) — deck-only, ephemeral
+  // (not in history, not on the overlay, gone on reload). Capped so hours of
+  // talking can't grow the DOM without bound.
+  const MAX_HEARD_ROWS = 30;
+  function addHeard(h) {
+    clearEmpty();
+    const rows = feed.querySelectorAll('.heard');
+    if (rows.length >= MAX_HEARD_ROWS) rows[0].remove();
+    const div = document.createElement('div');
+    div.className = 'heard' + (h.channel === 'party' ? ' party' : '');
+    div.title = T('heardTip');
+    const time = new Date(h.ts).toLocaleTimeString([], { hour: '2-digit', minute: '2-digit' });
+    div.innerHTML = `<span class="ic">${h.channel === 'party' ? '🎧' : '🎙️'}</span><span class="txt"></span><span class="tm">${time}</span>`;
+    div.querySelector('.txt').textContent = (h.channel === 'party' && h.label ? h.label + ' · ' : '') + '“' + h.text + '”';
+    feed.appendChild(div);
+    feed.scrollTop = feed.scrollHeight;
+  }
+
   function addSystem(text) {
     clearEmpty();
     const div = document.createElement('div');
@@ -537,6 +565,7 @@
       else if (msg.type === 'moment') { momentsData.push(msg.moment); addMoment(msg.moment, true); }
       else if (msg.type === 'state') { const s = state._startedAt; state = msg.state; state._startedAt = state.startedAt || s || Date.now(); renderState(); }
       else if (msg.type === 'system') { addSystem(msg.text); }
+      else if (msg.type === 'heard') { addHeard(msg.heard); }
     };
     ws.onclose = () => { if (!announcedDrop) { announcedDrop = true; toast(T('lost')); } setTimeout(connect, 3000); };
   }

public/settings.html

@@ -212,6 +212,10 @@ <h2>🎙️ Voice awareness</h2>
           <div><label>Recent speech window (seconds)</label><input type="number" data-path="cadence.transcriptWindowSeconds" min="30" max="600"></div>
           <div><label>Check for new speech every (seconds)</label><input type="number" data-path="transcript.pollSeconds" min="1" max="30"></div>
         </div>
+        <div class="check">
+          <input type="checkbox" id="showHeard" data-path="transcript.showInFeed" data-type="bool">
+          <label for="showHeard">Show what it hears in the deck feed — faint 🎙 lines of your transcribed speech (and 🎧 party audio), so you can spot mishears the cast might react to. Deck only; never on the stream overlay. Applies live on save.</label>
+        </div>
 
         <div style="margin-top:24px">
           <h2>🎧 Second channel — party / co-op audio</h2>

src/server.js

@@ -234,7 +234,7 @@ export function startServer(opts = {}) {
   const HOT_PATHS = [
     'energy', 'talkingPoints', 'theme.', 'overlay.', 'moments.', 'memory.', 'stream.',
     'app.costMeter', 'app.uiLanguage', 'app.fontScale', 'app.autoPause', 'app.autoPauseMinutes', 'app.autoUpdate',
-    'app.ttsVoice', 'app.ttsRate', 'app.ttsOutputDevice',
+    'app.ttsVoice', 'app.ttsRate', 'app.ttsOutputDevice', 'transcript.showInFeed',
     'cadence.soloSeconds', 'cadence.quietSeconds', 'cadence.jitter', 'cadence.burstChance',
     'cadence.lullChance', 'cadence.replyDelaySeconds', 'cadence.minVoiceReplyGapSeconds',
     'cadence.minScreenshotGapSeconds', 'cadence.minPartyNudgeGapSeconds',
@@ -556,6 +556,13 @@ export function startServer(opts = {}) {
     obs,
     onSpeech: (line) => {
       state.lastHeard = { text: line, ts: Date.now() };
+      // Deck-only echo of what was transcribed, so mishears are visible the
+      // moment they happen. Never enters `history` (the brain already gets
+      // the transcript window separately) and never reaches the overlay
+      // (its client ignores the type). Read live — the toggle hot-applies.
+      if (config.transcript.showInFeed !== false) {
+        broadcast({ type: 'heard', heard: { channel: 'mic', text: line, ts: Date.now() } });
+      }
       broadcastState();
       loop.onSpeech();
     },
@@ -571,6 +578,13 @@ export function startServer(opts = {}) {
     obs,
     onSpeech: (line) => {
       state.lastHeardParty = { text: line, ts: Date.now() };
+      // Same deck-only echo as the mic channel, labeled as "other people".
+      if (config.transcript.showInFeed !== false) {
+        broadcast({
+          type: 'heard',
+          heard: { channel: 'party', label: config.transcript2?.label || 'Party', text: line, ts: Date.now() },
+        });
+      }
       broadcastState();
       loop.onPartySpeech();
     },

test/transcript.test.js

@@ -88,3 +88,33 @@ test('getWindow prunes entries older than the window and honors sinceTs', (t) =>
   assert.equal(feed.getWindow(), 'new speech');
   assert.equal(feed.getWindow(Date.now() + 1000), ''); // nothing newer than the future
 });
+
+// textSource mode must drive the exact same onSpeech path as file mode — the
+// deck's "heard" feed echo and voice replies hang off that callback, so both
+// transcription modes get identical behavior.
+test('textSource mode fires onSpeech on changed captions, same as file mode', async (t) => {
+  let sourceText = 'stale pre-launch caption';
+  const heard = [];
+  const feed = new TranscriptFeed({
+    mode: 'textSource',
+    textSource: 'LocalVocal Captions',
+    pollSeconds: 999,
+    windowSeconds: 120,
+    obs: { getTextSourceText: async () => sourceText },
+    onSpeech: (line) => heard.push(line),
+  });
+  // Prime exactly like start() does: the pre-launch caption is not fresh speech.
+  feed.lastSourceText = await feed.obs.getTextSourceText(feed.textSource);
+  await feed.pollTextSource();
+  assert.deepEqual(heard, [], 'primed caption must not fire');
+  sourceText = 'did you see that dragon';
+  await feed.pollTextSource();
+  await feed.pollTextSource(); // unchanged caption fires only once
+  assert.deepEqual(heard, ['did you see that dragon']);
+  sourceText = null; // OBS unreachable mid-stream
+  await feed.pollTextSource();
+  sourceText = '00:01:02,000 --> 00:01:04,000'; // SRT timing junk is filtered here too
+  await feed.pollTextSource();
+  assert.deepEqual(heard, ['did you see that dragon']);
+  assert.equal(feed.getWindow(), 'did you see that dragon');
+});