diff --git a/CHANGELOG.md b/CHANGELOG.md
index 4ba9912..2e9093a 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -7,105 +7,75 @@ Follows [Keep a Changelog](https://keepachangelog.com/en/1.0.0/) · versioned wi
 
 ## [Unreleased]
 
-### Performance
+---
 
-- **Self-hosted WaveSurfer and MDI icon font** — removed two render-blocking
-  `<script>` and one blocking `<link>` to `unpkg.com` / `cdn.jsdelivr.net`
-  from `<head>`. WaveSurfer (~57 KB) and MDI font (~394 KB WOFF2 + CSS) are
-  now served locally from `static/vendor/`. Flag-icons CSS is loaded async
-  (non-blocking) via the `rel=preload` / `onload` trick.
-- **Skeleton loading view** — `index.html` shows an animated shimmer
-  placeholder (page head + toolbar + 8 voice card outlines) immediately on
-  first paint, before any JS loads. The skeleton fades out once `loader.js`
-  finishes and real sections are revealed. No external deps — pure inline CSS.
-
-### Fixed
-
-- **LLM returned empty response (Qwen3 thinking mode)** — Qwen3 models
-  stream thinking tokens under `delta.reasoning_content` instead of
-  `delta.content`. The conversation turn now falls back to
-  `reasoning_content` so think-only responses produce visible output.
-  Error message improved with a hint to add `/no-think` to the system prompt.
-
-### Performance
-
-- **Parallel JS module loading** — `loader.js` previously loaded all 16
-  modules sequentially (17 round-trips). Restructured into 4 ordered batches
-  with `async=false` so files are fetched in parallel but execute in the
-  correct dependency order:
-  `utils` → `settings` → *(9 feature modules in parallel)* → `init` →
-  *(4 post-init modules in parallel)* → `nav`
-  Round-trips reduced from 17 to 6; 9 files now download simultaneously.
-- **Version-based JS/CSS cache busting** — `loader.js` fetches
-  `/api/version` first and appends `?v=<version>` to every script URL.
-  A new `Cache-Control: public, max-age=31536000, immutable` middleware in
-  `server.py` lets the browser cache versioned assets for a full year.
-  Bumping the version (via `scripts/release.py`) invalidates the cache.
-  Section HTML keeps `?v=<timestamp>` (no-store) so it is always fresh.
-
-### Changed
-
-- **Config and logs are now bind-mounted local folders** — replaced the
-  opaque named Docker volume (`tts-voice-creator-clone-and-design-2`) with
-  two transparent host directories:
-  - `./config/` → `/home/app/.config/tts-voice-creator` — holds
-    `settings.json`, `voice_design_presets.json`, `tts_routes.json`
-  - `./logs/`   → `/logs` — holds `app.log` (rotates at 5 MB, 3 backups)
-  Both directories are tracked in git (via `.gitkeep`) but their runtime
-  contents are excluded from version control via `.gitignore`.
-  `portainer-stack.yml` updated with absolute host paths.
-- **Server writes a rotating log file** — `RotatingFileHandler` added to
-  `server.py`; writes `INFO`-level and above to `./logs/app.log`.
-
-### Fixed
-
-- **Engine settings lost after container recreate** — container names and
-  dynamic Docker card URL overrides were stored only in `localStorage`.
-  Added `engine_container_names` as a persisted server setting; all
-  container name inputs tagged with `data-cn-key` so `loadSettings()`
-  can restore them; dynamic card URL overrides now saved under
-  `engine_local_urls` with a `dc-` prefix. On first page load the server
-  wins over `localStorage`; changes write to both immediately.
-
-- **`ImportError: cannot import name '_AUDIO_EXTS' from 'core.audio'`** —
-  `routes/stt.py` had a dead alias `from core.audio import _to_wav_16k, _AUDIO_EXTS as _VOICE_AUDIO_EXTS`;
-  `_AUDIO_EXTS` lives in `core.voice`, not `core.audio`. Removed the
-  wrong import; `_AUDIO_EXTS` is already correctly imported from `core.voice`
-  on the next line and the unused alias was never referenced in the file.
-
-- **`No module named 'core'` on container restart** — `docker restart` reuses
-  the old container configuration and never applies new volume mounts from
-  `docker-compose.yml`. Added a comment to `portainer-stack.yml` and updated
-  it to include `core:/app/core:ro` and `routes:/app/routes:ro`. Solution is
-  `docker compose up -d` (recreates the container) not `docker restart`.
-
-- **Text-input turns returned 422** — `UploadFile | None = File(None)` with
-  `from __future__ import annotations` caused FastAPI to treat `audio` as a
-  required field even when omitted. Changed to `Optional[UploadFile] = None`
-  (no `File()` wrapper) so the field is genuinely optional for text-only turns.
-
-- **Conversation input bar hidden when mic unavailable** — the mic-blocked
-  warning box was inserted before `conv-chat-window` inside the flex column,
-  pushing the input bar off-screen. Fixed by prepending the warning inside
-  `conv-chat-window` so it scrolls with the chat and never affects the bar.
-- **Chat panel overflow / missing input bar on smaller viewports** — removed
-  `min-height: 400px` from `conv-chat-window` (prevented shrinking) and gave
-  `conv-chat-panel` a viewport-relative height (`calc(100vh - 340px)`,
-  min 440px) so the input bar is always anchored at the bottom.
-- **Browser caches old section HTML after updates** — `loader.js` now appends
-  `?v=<timestamp>` to every section fetch, busting the cache on each page load.
+## [1.2.0] — 2026-05-29
 
 ### Added
 
+- **Remember last section on reload** — the active section (and Settings /
+  Engines sub-page) is persisted in `localStorage`. A hard-reload
+  (Ctrl+Shift+R) now returns to the same page instead of always jumping to
+  My Voices.
+- **Conversation: live speech preview** — while recording, the active Whisper
+  STT backend transcribes accumulated audio every 2.5 s and shows the result
+  in the text input field in real time. The input is pre-populated with this
+  live guess before the final Whisper result arrives. Also tries the browser's
+  Web Speech API first (works on HTTPS / localhost) for even faster results.
+- **Conversation: Voice Activity Detection (VAD)** — recording now auto-stops
+  after 1.5 s of silence detected via the Web Audio `AnalyserNode` RMS level.
+  A "Sending in X.Xs" countdown appears in the status bar so the timing is
+  visible. An **Auto-stop** toggle in the input bar lets users disable VAD and
+  revert to click-to-stop. A thin audio-level bar below the status line shows
+  microphone volume in real time during recording.
+- **Conversation: hands-free mode** — after the agent finishes speaking, the
+  microphone restarts automatically. A **Hands-free** toggle (on by default)
+  disables this; clicking the mic button manually always cancels any pending
+  auto-restart.
 - **`scripts/release.py`** — automates version bump + CHANGELOG promotion.
   `python scripts/release.py --patch|--minor|--major [--dry-run]` renames
   `[Unreleased]` to the new version, updates compare links, writes `VERSION`,
   commits, and creates an annotated git tag in one command.
-- **Git pre-commit hook** (`scripts/hooks/pre-commit`) — warns (does not block)
-  when `.py`/`.js`/`.css`/`.html` files are staged but `CHANGELOG.md` or
-  `VERSION` are not. Run `bash scripts/install-hooks.sh` after cloning.
-- **`scripts/install-hooks.sh`** — one-liner to install the hook after a fresh
-  clone: `bash scripts/install-hooks.sh`.
+- **Git pre-commit hook** (`scripts/hooks/pre-commit`) — warns (does not
+  block) when `.py`/`.js`/`.css`/`.html` files are staged but `CHANGELOG.md`
+  or `VERSION` are not. Run `bash scripts/install-hooks.sh` after cloning.
+- **`scripts/install-hooks.sh`** — one-liner to install the hook after a
+  fresh clone: `bash scripts/install-hooks.sh`.
+
+### Changed
+
+- **Config and logs are now bind-mounted local folders** — replaced the
+  opaque named Docker volume with `./config/` and `./logs/` host directories.
+  `portainer-stack.yml` updated with absolute host paths.
+- **Server writes a rotating log file** — `RotatingFileHandler` writes
+  `INFO`-level and above to `./logs/app.log` (rotates at 5 MB, 3 backups).
+
+### Performance
+
+- **Skeleton loading view** — `index.html` shows an animated shimmer
+  placeholder immediately on first paint; fades out once JS finishes loading.
+- **Self-hosted WaveSurfer and MDI icon font** — removed render-blocking CDN
+  requests; assets now served locally from `static/vendor/`.
+- **Parallel JS module loading** — restructured `loader.js` into 4 ordered
+  batches; round-trips reduced from 17 to 6, 9 files fetched simultaneously.
+- **Version-based JS/CSS cache busting** — versioned assets served with
+  `max-age=31536000, immutable`; bumping version invalidates the cache.
+
+### Fixed
+
+- **LLM returned empty response (Qwen3 thinking mode)** — conversation turn
+  now falls back to `reasoning_content` for think-only responses; error
+  message hints to add `/no-think` to the system prompt.
+- **Engine settings lost after container recreate** — container names and URL
+  overrides now persisted as server settings (`engine_container_names`,
+  `engine_local_urls`); restored from server on first page load.
+- **Text-input turns returned 422** — changed `audio` form field to
+  `Optional[UploadFile] = None` so text-only turns don't require audio.
+- **Conversation input bar hidden when mic unavailable** — warning box moved
+  inside `conv-chat-window` so it never pushes the input bar off-screen.
+- **Browser caches old section HTML** — `loader.js` appends `?v=<timestamp>`
+  to every section fetch.
+- Various import errors and container restart issues fixed.
 
 ---
 
@@ -279,6 +249,7 @@ Initial feature-complete release.
 
 ---
 
-[Unreleased]: https://github.com/mARTin-B78/tts-voice-creator-clone-and-design-2/compare/v1.1.0...HEAD
+[Unreleased]: https://github.com/mARTin-B78/tts-voice-creator-clone-and-design-2/compare/v1.2.0...HEAD
+[1.2.0]: https://github.com/mARTin-B78/tts-voice-creator-clone-and-design-2/compare/v1.1.0...v1.2.0
 [1.1.0]: https://github.com/mARTin-B78/tts-voice-creator-clone-and-design-2/compare/v1.0.0...v1.1.0
 [1.0.0]: https://github.com/mARTin-B78/tts-voice-creator-clone-and-design-2/releases/tag/v1.0.0
diff --git a/VERSION b/VERSION
index 9084fa2..26aaba0 100644
--- a/VERSION
+++ b/VERSION
@@ -1 +1 @@
-1.1.0
+1.2.0
diff --git a/static/js/conversation.js b/static/js/conversation.js
index e6f1514..276faef 100644
--- a/static/js/conversation.js
+++ b/static/js/conversation.js
@@ -173,6 +173,19 @@ $('s-import-voices-file')?.addEventListener('change', async function () {
   let conversationHistory = [];
   let turnCount = 0;
   let isProcessing = false;
+  let vadAudioCtx = null;
+  let vadRafId = null;
+  let liveInterimText = '';
+  let speechRec = null;
+  const VAD_THRESHOLD = 0.01;
+  const VAD_MIN_REC_MS = 500;
+  const VAD_SILENCE_MS = 1500;
+  const origPlaceholder = textInput?.placeholder || '';
+  const vadToggle       = $('conv-vad-toggle');
+  const handsFreeToggle = $('conv-handsfree-toggle');
+  let previewTranscribing = false;
+  let convCurrentAudio = null;
+  let autoMicGeneration = 0;
 
   // ── Populate STT backends ────────────────────────────────────────────────
   async function loadConvSttBackends() {
@@ -363,6 +376,45 @@ $('s-import-voices-file')?.addEventListener('change', async function () {
     if (micTimer) micTimer.textContent = '';
   }
 
+  // ── Chunked Whisper preview (fallback for browsers without SpeechRecognition) ──
+  async function transcribeForPreview() {
+    if (previewTranscribing || !recChunks.length) return;
+    previewTranscribing = true;
+    try {
+      const mime = (mediaRecorder && mediaRecorder.mimeType) || 'audio/webm';
+      const ext = mime.includes('ogg') ? '.ogg' : '.webm';
+      const blob = new Blob(recChunks, { type: mime });
+      const fd = new FormData();
+      fd.append('file', new File([blob], 'preview' + ext, { type: mime }));
+      fd.append('backend', sttSel?.value || 'configured');
+      const r = await fetch('/api/transcribe-bytes', { method: 'POST', body: fd });
+      if (r.ok) {
+        const d = await r.json();
+        const txt = (d.text || '').trim();
+        if (txt && mediaRecorder && mediaRecorder.state === 'recording') {
+          liveInterimText = txt;
+          if (textInput) textInput.value = txt;
+        }
+      }
+    } catch (_) {}
+    finally { previewTranscribing = false; }
+  }
+
+  // ── Auto-restart mic after agent finishes speaking ───────────────────────
+  function scheduleAutoMic() {
+    if (!handsFreeToggle?.checked) return;
+    const gen = ++autoMicGeneration;
+    function tryStart() {
+      if (gen !== autoMicGeneration || isProcessing) return;
+      startRecording().catch(() => {});
+    }
+    if (convCurrentAudio && !convCurrentAudio.ended) {
+      convCurrentAudio.addEventListener('ended', () => setTimeout(tryStart, 200), { once: true });
+    } else {
+      setTimeout(tryStart, 300);
+    }
+  }
+
   async function startRecording() {
     if (isProcessing) return;
     if (!navigator.mediaDevices?.getUserMedia) {
@@ -377,23 +429,119 @@ $('s-import-voices-file')?.addEventListener('change', async function () {
       toast('Microphone access denied: ' + e.message, 'error');
       return;
     }
+
+    liveInterimText = '';
+    previewTranscribing = false;
     recChunks = [];
     mediaRecorder = new MediaRecorder(stream);
-    mediaRecorder.ondataavailable = e => { if (e.data.size > 0) recChunks.push(e.data); };
+    mediaRecorder.ondataavailable = e => {
+      if (e.data.size > 0) {
+        recChunks.push(e.data);
+        // Trigger Whisper preview on periodic chunks (not on the final stop chunk)
+        if (mediaRecorder.state === 'recording') transcribeForPreview();
+      }
+    };
     mediaRecorder.onstop = () => {
       stream.getTracks().forEach(t => t.stop());
       const blob = new Blob(recChunks, { type: mediaRecorder.mimeType || 'audio/webm' });
       processBlob(blob);
     };
-    mediaRecorder.start();
+    // timeslice=2500: ondataavailable fires every 2.5 s → interim Whisper preview
+    mediaRecorder.start(2500);
     micBtn.classList.add('recording');
     micIcon.className = 'mdi mdi-stop';
-    if (micStatus) micStatus.textContent = 'Recording… click to stop';
+    if (micStatus) micStatus.textContent = 'Recording…';
     startRecTimer();
+
+    // Set input to listening mode (read-only; Whisper preview text will appear here)
+    if (textInput) {
+      textInput.readOnly = true;
+      textInput.value = '';
+      textInput.placeholder = 'Listening…';
+      textInput.classList.add('listening');
+    }
+    if (sendBtn) sendBtn.disabled = true;
+
+    // Also try Web Speech API for faster interim results (works on HTTPS / localhost)
+    const SpeechRec = window.SpeechRecognition || window.webkitSpeechRecognition;
+    if (SpeechRec) {
+      try {
+        speechRec = new SpeechRec();
+        speechRec.continuous = true;
+        speechRec.interimResults = true;
+        speechRec.onresult = e => {
+          let final = '', interim = '';
+          for (let i = 0; i < e.results.length; i++) {
+            if (e.results[i].isFinal) final += e.results[i][0].transcript;
+            else interim += e.results[i][0].transcript;
+          }
+          liveInterimText = final + interim;
+          if (textInput) textInput.value = liveInterimText;
+        };
+        speechRec.onerror = () => { speechRec = null; };
+        speechRec.start();
+      } catch(_) { speechRec = null; }
+    }
+
+    // VAD: auto-stop on silence using AudioContext
+    if (vadToggle?.checked) {
+      try {
+        vadAudioCtx = new AudioContext();
+        const src = vadAudioCtx.createMediaStreamSource(stream);
+        const analyser = vadAudioCtx.createAnalyser();
+        analyser.fftSize = 1024;
+        src.connect(analyser);
+        const vadBuf = new Float32Array(analyser.fftSize);
+        let vadLastVoice = Date.now();
+        const levelWrap = $('conv-level-wrap');
+        const levelFill = $('conv-level-fill');
+        if (levelWrap) levelWrap.classList.add('active');
+
+        function vadTick() {
+          if (!mediaRecorder || mediaRecorder.state !== 'recording') return;
+          analyser.getFloatTimeDomainData(vadBuf);
+          let rms = 0;
+          for (const s of vadBuf) rms += s * s;
+          rms = Math.sqrt(rms / vadBuf.length);
+          if (levelFill) levelFill.style.width = Math.min(100, rms * 5000) + '%';
+          if (rms > VAD_THRESHOLD) vadLastVoice = Date.now();
+          const elapsed = Date.now() - recStart;
+          const silence = Date.now() - vadLastVoice;
+          if (elapsed > VAD_MIN_REC_MS) {
+            const remaining = VAD_SILENCE_MS - silence;
+            if (micStatus) {
+              micStatus.textContent = remaining < VAD_SILENCE_MS * 0.7
+                ? `Sending in ${(Math.max(0, remaining) / 1000).toFixed(1)}s…`
+                : 'Recording…';
+            }
+            if (silence >= VAD_SILENCE_MS) {
+              stopRecording();
+              return;
+            }
+          }
+          vadRafId = requestAnimationFrame(vadTick);
+        }
+        vadRafId = requestAnimationFrame(vadTick);
+      } catch(_) { vadAudioCtx = null; }
+    }
   }
 
   function stopRecording() {
     if (!mediaRecorder || mediaRecorder.state === 'inactive') return;
+
+    // Clean up VAD
+    cancelAnimationFrame(vadRafId);
+    vadRafId = null;
+    if (vadAudioCtx) { try { vadAudioCtx.close(); } catch(_){} vadAudioCtx = null; }
+    const levelWrap = $('conv-level-wrap');
+    const levelFill = $('conv-level-fill');
+    if (levelWrap) levelWrap.classList.remove('active');
+    if (levelFill) levelFill.style.width = '0%';
+
+    // Clean up live speech recognition
+    if (speechRec) { try { speechRec.stop(); } catch(_){} speechRec = null; }
+    previewTranscribing = false;
+
     mediaRecorder.stop();
     stopRecTimer();
     micBtn.classList.remove('recording');
@@ -405,12 +553,24 @@ $('s-import-voices-file')?.addEventListener('change', async function () {
 
   // ── Send turn via SSE ─────────────────────────────────────────────────────
   async function processBlob(blob) {
+    autoMicGeneration++; // cancel any pending auto-mic from previous turn
+    convCurrentAudio = null;
     turnCount++;
     const turnN = turnCount;
     const t0 = Date.now();
 
-    // Show user bubble with placeholder
-    const userBubble = addBubble('user', '…');
+    // Transition input from listening/readOnly state to processing/disabled state
+    if (textInput) {
+      textInput.readOnly = false;
+      textInput.classList.remove('listening');
+      textInput.placeholder = origPlaceholder;
+      textInput.value = '';
+      textInput.disabled = true;
+    }
+    if (sendBtn) sendBtn.disabled = true;
+
+    // Show user bubble seeded with live interim transcript (if available)
+    const userBubble = addBubble('user', liveInterimText || '…');
     const assistantBubble = addTypingBubble();
     let assistantText = '';
     let lastStats = null;
@@ -464,9 +624,10 @@ $('s-import-voices-file')?.addEventListener('change', async function () {
             for (let i = 0; i < binStr.length; i++) arr[i] = binStr.charCodeAt(i);
             const audioBlob = new Blob([arr], { type: mime });
             const url = URL.createObjectURL(audioBlob);
-            const audio = new Audio(url);
-            audio.onended = () => URL.revokeObjectURL(url);
-            audio.play().catch(() => {});
+            const audioEl = new Audio(url);
+            convCurrentAudio = audioEl;
+            audioEl.addEventListener('ended', () => { URL.revokeObjectURL(url); convCurrentAudio = null; });
+            audioEl.play().catch(() => {});
             if (micStatus) micStatus.textContent = 'Speaking…';
           } else if (evt.type === 'stats') {
             lastStats = evt;
@@ -475,7 +636,8 @@ $('s-import-voices-file')?.addEventListener('change', async function () {
             conversationHistory.push({ role: 'user', content: userBubble.textContent });
             conversationHistory.push({ role: 'assistant', content: assistantText });
             addHistoryItem(turnN, lastStats?.total_ms ?? (Date.now() - t0), true);
-            if (micStatus) micStatus.textContent = 'Ready';
+            scheduleAutoMic();
+            if (!handsFreeToggle?.checked && micStatus) micStatus.textContent = 'Ready';
           } else if (evt.type === 'error') {
             const wrap = assistantBubble.closest('.conv-bubble-wrap');
             if (wrap) wrap.remove();
@@ -499,11 +661,13 @@ $('s-import-voices-file')?.addEventListener('change', async function () {
       addHistoryItem(turnN, Date.now() - t0, false);
       if (micStatus) micStatus.textContent = 'Error — ready';
     } finally {
+      liveInterimText = '';
+      previewTranscribing = false;
       isProcessing = false;
       micBtn.classList.remove('processing');
       micIcon.className = 'mdi mdi-microphone';
       if (sendBtn) sendBtn.disabled = false;
-      if (textInput) textInput.disabled = false;
+      if (textInput) { textInput.disabled = false; textInput.value = ''; }
     }
   }
 
@@ -511,6 +675,8 @@ $('s-import-voices-file')?.addEventListener('change', async function () {
   async function processText(text) {
     text = text.trim();
     if (!text || isProcessing) return;
+    autoMicGeneration++; // cancel any pending auto-mic
+    convCurrentAudio = null;
     isProcessing = true;
     if (sendBtn) sendBtn.disabled = true;
     if (textInput) { textInput.disabled = true; textInput.value = ''; }
@@ -573,9 +739,10 @@ $('s-import-voices-file')?.addEventListener('change', async function () {
             for (let i = 0; i < binStr.length; i++) arr[i] = binStr.charCodeAt(i);
             const audioBlob = new Blob([arr], { type: mime });
             const url = URL.createObjectURL(audioBlob);
-            const audio = new Audio(url);
-            audio.onended = () => URL.revokeObjectURL(url);
-            audio.play().catch(() => {});
+            const audioEl = new Audio(url);
+            convCurrentAudio = audioEl;
+            audioEl.addEventListener('ended', () => { URL.revokeObjectURL(url); convCurrentAudio = null; });
+            audioEl.play().catch(() => {});
             if (micStatus) micStatus.textContent = 'Speaking…';
           } else if (evt.type === 'stats') {
             lastStats = evt;
@@ -584,7 +751,8 @@ $('s-import-voices-file')?.addEventListener('change', async function () {
             conversationHistory.push({ role: 'user', content: text });
             conversationHistory.push({ role: 'assistant', content: assistantText });
             addHistoryItem(turnN, lastStats?.total_ms ?? (Date.now() - t0), true);
-            if (micStatus) micStatus.textContent = 'Ready';
+            scheduleAutoMic();
+            if (!handsFreeToggle?.checked && micStatus) micStatus.textContent = 'Ready';
           } else if (evt.type === 'error') {
             const wrap = assistantBubble.closest('.conv-bubble-wrap');
             if (wrap) wrap.remove();
@@ -612,6 +780,7 @@ $('s-import-voices-file')?.addEventListener('change', async function () {
   // ── Wire up events ───────────────────────────────────────────────────────
   micBtn.addEventListener('click', () => {
     if (isProcessing) return;
+    autoMicGeneration++; // cancel any pending hands-free auto-restart
     if (mediaRecorder && mediaRecorder.state === 'recording') {
       stopRecording();
     } else {
diff --git a/static/nav.js b/static/nav.js
index 5de071b..702c8db 100644
--- a/static/nav.js
+++ b/static/nav.js
@@ -40,6 +40,7 @@
   }
 
   function showSection(sectionId) {
+    localStorage.setItem('ttsvc_section', sectionId);
     SECTIONS.forEach(function (id) {
       var el = document.getElementById(id);
       if (el) el.classList.toggle('is-active', id === sectionId);
@@ -124,6 +125,7 @@
 
   window.navSettingsCat = function (cat) {
     window._settingsSidebarCat = cat;
+    localStorage.setItem('ttsvc_settings_cat', cat);
     navTo('s-settings');   // showSection will call applySettingsPage(cat)
     if (cat === 'logs' && typeof loadSettingsLogs === 'function') loadSettingsLogs();
     if (cat === 'about' && typeof renderSettingsAbout === 'function') renderSettingsAbout();
@@ -143,11 +145,17 @@
 
   window.navEnginesCat = function (cat) {
     window._enginesSidebarCat = cat;
+    localStorage.setItem('ttsvc_engines_cat', cat);
     navTo('s-llms');   // showSection will call applyEnginesPage(cat)
     if ((cat === 'tts' || cat === 'stt') && typeof loadLocalContainers === 'function') loadLocalContainers();
   };
 
-  // Show default section on load
-  showSection('s-voices');
-  runSideEffects('library');
+  // Restore last-used section from localStorage (fallback: My Voices)
+  var _savedSection = localStorage.getItem('ttsvc_section');
+  window._settingsSidebarCat = localStorage.getItem('ttsvc_settings_cat') || 'connections';
+  window._enginesSidebarCat  = localStorage.getItem('ttsvc_engines_cat')  || 'llm';
+  var _startSection = (_savedSection && SECTIONS.includes(_savedSection)) ? _savedSection : 's-voices';
+  showSection(_startSection);
+  var _startTab = Object.keys(TAB_SECTION_MAP).find(function (k) { return TAB_SECTION_MAP[k] === _startSection; }) || 'library';
+  runSideEffects(_startTab);
 })();
diff --git a/static/sections/s-conversation.html b/static/sections/s-conversation.html
index 1fcea0f..e6102ce 100644
--- a/static/sections/s-conversation.html
+++ b/static/sections/s-conversation.html
@@ -54,6 +54,7 @@
           <!-- Input bar: text field + send + mic -->
           <div class="conv-input-bar">
             <div class="conv-mic-status" id="conv-mic-status">Ready</div>
+            <div class="conv-level-wrap" id="conv-level-wrap"><div class="conv-level-fill" id="conv-level-fill"></div></div>
             <div class="conv-text-row">
               <input id="conv-text-input" class="conv-text-inp"
                 type="text" placeholder="Type a message and press Enter or →"
@@ -66,6 +67,14 @@
                 <span class="mdi mdi-microphone" id="conv-mic-icon"></span>
               </button>
               <div class="conv-mic-timer" id="conv-mic-timer"></div>
+              <label class="conv-vad-label" title="Auto-stop recording when silence is detected">
+                <input type="checkbox" id="conv-vad-toggle" checked>
+                <span>Auto-stop</span>
+              </label>
+              <label class="conv-vad-label" title="Automatically restart microphone after the agent finishes speaking">
+                <input type="checkbox" id="conv-handsfree-toggle" checked>
+                <span>Hands-free</span>
+              </label>
             </div>
           </div>
         </div>
diff --git a/static/style.css b/static/style.css
index 56afb7d..54ff761 100644
--- a/static/style.css
+++ b/static/style.css
@@ -2093,6 +2093,12 @@ code { background: var(--panel); border-radius: 4px; padding: 1px 5px; font-fami
 .conv-mic-btn.processing { background: var(--yellow); border-color: var(--yellow); cursor: default; }
 @keyframes convPulse { 0%,100% { box-shadow: 0 0 0 0 rgba(220,38,38,.4); } 50% { box-shadow: 0 0 0 8px rgba(220,38,38,0); } }
 .conv-mic-timer { font-size: 13px; font-variant-numeric: tabular-nums; color: var(--red); min-width: 36px; text-align: right; }
+.conv-vad-label { display: flex; align-items: center; gap: 4px; font-size: 11px; color: var(--subtext); cursor: pointer; white-space: nowrap; user-select: none; }
+.conv-vad-label input { cursor: pointer; accent-color: var(--accent); }
+.conv-level-wrap { height: 2px; background: var(--panel); border-radius: 1px; margin: 3px 0; overflow: hidden; display: none; }
+.conv-level-wrap.active { display: block; }
+.conv-level-fill { height: 100%; background: var(--red); border-radius: 1px; width: 0%; }
+.conv-text-inp.listening { border-color: var(--red); font-style: italic; color: var(--subtext); }
 
 /* Stats panel */
 .conv-stats-head { font-size: 10px; font-weight: 800; color: var(--subtext); text-transform: uppercase; letter-spacing: .08em; }
