diff --git a/simulator/e2e/tests/altitude-lock.spec.ts b/simulator/e2e/tests/altitude-lock.spec.ts
index 4c9c3d9..a6da1d3 100644
--- a/simulator/e2e/tests/altitude-lock.spec.ts
+++ b/simulator/e2e/tests/altitude-lock.spec.ts
@@ -83,13 +83,14 @@ test("audio is a 0-10 level dial", async ({ page }) => {
   expect(ctl.max).toBe("10");
 });
 
-test("the experience auto-starts once loaded: video on + audio at 3/10", async ({ page }) => {
+test("boots silent (video off, audio 0); turning video on couples audio to 3/10", async ({ page }) => {
   await boot(page);
-  // No gesture needed — autoStart() flips Video on and lifts audio to a gentle level
-  // the moment the preload finishes.
-  await expect.poll(() => page.evaluate(() => (document.querySelector("#visual") as HTMLInputElement).checked)).toBe(true);
+  // Boots SILENT — no un-gestured auto-start, so the browser autoplay block never bites.
+  expect(await page.evaluate(() => (document.querySelector("#visual") as HTMLInputElement).checked)).toBe(false);
+  expect(await page.evaluate(() => (document.querySelector("#audio") as HTMLInputElement).value)).toBe("0");
+  // Turning Video on (a click gesture) lifts the audio dial to a gentle 3/10.
+  await enableVideo(page);
   await expect.poll(() => page.evaluate(() => (document.querySelector("#audio") as HTMLInputElement).value)).toBe("3");
-  // The level label reflects it.
   expect(await page.evaluate(() => document.querySelector("#audio-level-val")?.textContent)).toBe("3");
 });
 
diff --git a/simulator/static/app.js b/simulator/static/app.js
index dbe005e..7fe28d1 100644
--- a/simulator/static/app.js
+++ b/simulator/static/app.js
@@ -1260,8 +1260,12 @@ async function main() {
   $("stage").addEventListener("wheel", onWheel, { passive: false });
   update();                   // render the initial state (both toggles off → black, silent)
   await preloadAllMedia();    // download all media, updating the loading bar
-  hideLoading();              // experience is ready to run smoothly
-  autoStart();                // once loaded: video on + gentle audio (3/10), no gesture needed
+  hideLoading();              // experience is ready — boots SILENT (video off, audio 0)
+  // No un-gestured auto-start: a browser autoplay policy blocks an audio play() made
+  // outside a user gesture (muted video survives, sound does not), so an auto-start
+  // would show video but stay silent. Instead the experience waits for the operator to
+  // turn Video on — that click IS the gesture, and its handler (above) lifts audio to
+  // 3/10 and starts the soundtrack in-gesture, so sound reliably comes up with video.
 }
 
 // Populate the language dropdown from the registry and wire live switching.
@@ -1297,16 +1301,4 @@ function setLanguage(lang) {
   if (lastAffect) renderAffect(lastAffect.strength, lastAffect.intensity, lastAffect.right);
 }
 
-// Auto-start the experience the moment media is ready: bring Video on and the audio
-// level to 3/10 (the same gentle default the first manual Video-on couples in). The
-// audio <audio> elements ideally unlock inside a user gesture (Safari); this fires
-// outside one, so a browser with a strict autoplay policy may BLOCK the first audio
-// play() until any interaction — the on-screen audio status readout will show it.
-function autoStart() {
-  $("visual").checked = true;
-  videoEverOn = true;
-  if (audioLevel() === 0) { $("audio").value = "3"; updateAudioLevelLabel(); }
-  applyAudio();
-  update();
-}
 main();
diff --git a/tests/test_e2e_audio.py b/tests/test_e2e_audio.py
index 106ce25..f5d7065 100644
--- a/tests/test_e2e_audio.py
+++ b/tests/test_e2e_audio.py
@@ -2,12 +2,15 @@
 cleanly when Playwright or its browser binary is absent.
 
 NOTE: headless engines relax BOTH autoplay and GPU compositing, so these assert
-the WIRING (auto-start on load, dial→play/volume, video-off blanking) — real
-Safari/iOS autoplay and real-GPU compositing still need a device by-ear/eye check.
+the WIRING (boot-silent, video-on couples audio, dial→play/volume, video-off
+blanking) — real Safari/iOS autoplay and real-GPU compositing still need a device
+by-ear/eye check.
 
-The experience AUTO-STARTS once the preload finishes: Video turns on and the audio
-level dial rises to 3/10 with no user gesture (app.js autoStart()). Audio is a 0–10
-range dial (not a checkbox); set it by writing `value` + firing an `input` event.
+The experience BOOTS SILENT — video off, audio dial at 0 — so nothing plays until a
+real user gesture. Turning Video ON (a click) couples the audio dial up to 3/10 and
+starts the soundtrack IN that gesture, sidestepping the browser autoplay block that
+swallowed an un-gestured auto-start. Audio is a 0–10 range dial (not a checkbox);
+set it by writing `value` + firing an `input` event.
 
 Install:  pip install -e '.[e2e]' && python -m playwright install chromium
 """
@@ -63,7 +66,7 @@ def page(app_url):
         pg = browser.new_page()
         pg.goto(app_url)
         # wait out the "Loading Universe…" splash (it overlays + blocks clicks); the
-        # auto-start fires as it dismisses.
+        # experience boots silent (video off, audio 0) as it dismisses.
         pg.wait_for_selector("#loading", state="detached", timeout=60000)
         yield pg
         browser.close()
@@ -84,10 +87,23 @@ def _set_audio(page, level):
     )
 
 
-def test_app_auto_starts_video_and_audio(page):
-    # Once the preload finishes the experience auto-starts: Video on + audio at 3/10,
-    # video showing (black hidden), a scale soundtrack playing — no tap-to-start wall.
-    assert page.is_checked("#visual") is True
+def test_app_boots_silent_video_off(page):
+    # The experience boots SILENT: Video off, audio dial at 0, screen black, both
+    # crossfade elements paused — nothing plays until a real user gesture.
+    assert page.is_checked("#visual") is False
+    assert page.evaluate("document.getElementById('audio').value") == "0"
+    assert page.evaluate("document.getElementById('audio-level-val').textContent") == "0"
+    page.wait_for_selector("#black:not(.hidden)")                      # screen blanked
+    assert page.evaluate(
+        "(() => { const a = document.getElementById('aud'), b = document.getElementById('aud-b');"
+        " return a.paused && b.paused; })()"
+    ) is True
+
+
+def test_video_on_couples_audio_to_three_and_plays(page):
+    # Turning Video on (a click gesture) lifts the audio dial to 3/10 and starts the
+    # soundtrack IN that gesture: video shows (black hidden), a scale soundtrack plays.
+    _toggle_visual(page)                                                # off -> on
     page.wait_for_function("document.getElementById('audio').value === '3'")
     page.wait_for_function("document.getElementById('audio-level-val').textContent === '3'")
     page.wait_for_selector("#black.hidden", state="attached")          # video showing
@@ -113,8 +129,11 @@ def test_audio_dial_zero_silences(page):
     assert page.evaluate("document.getElementById('audio').value") == "0"
 
 
-def test_video_off_blanks_after_autostart(page):
-    # Video is auto-on; turning it off blanks the screen and hides the GPU layers.
+def test_video_off_blanks(page):
+    # Turn Video on (boot is off), then off again: that blanks the screen and hides
+    # the GPU layers.
+    _toggle_visual(page)                                                # off -> on
+    page.wait_for_selector("#black.hidden", state="attached")
     _toggle_visual(page)                                                # on -> off
     page.wait_for_selector("#black:not(.hidden)")
     page.wait_for_function(