diff --git a/wiki/SUMMARY.md b/wiki/SUMMARY.md
index 92e1f133..5b2925a6 100644
--- a/wiki/SUMMARY.md
+++ b/wiki/SUMMARY.md
@@ -146,6 +146,7 @@
   - [2026-07-22 Co GUI teaching-lock follow-up](playtests/2026-07-22-playtest-co-gui-teaching-lock.md)
   - [2026-07-24 Fable agent headless playtest](playtests/2026-07-24-playtest-fable-agent-headless.md)
   - [2026-07-27 Void agent opening playtest](playtests/2026-07-27-playtest-void-agent.md)
+  - [2026-07-27 Void GUI opening playtest](playtests/2026-07-27-playtest-void-gui-opening.md)
 
 # Log
 
diff --git a/wiki/playtests/2026-07-27-playtest-void-gui-opening.md b/wiki/playtests/2026-07-27-playtest-void-gui-opening.md
new file mode 100644
index 00000000..95d39a2a
--- /dev/null
+++ b/wiki/playtests/2026-07-27-playtest-void-gui-opening.md
@@ -0,0 +1,41 @@
+# 2026-07-27 — Playtest (Void): GUI opening on a virtual desktop
+
+```
+Type: log
+```
+
+This is a narrow GUI witness, not a full playthrough. Cameron asked Void to run the graphical frontend directly rather than use the agent-playtester workflow. I launched the real Bevy window from `main` at `fbf064ed0` inside a 1280×720 X11 virtual desktop with Openbox, focused it, sent keyboard input, and captured full-window screenshots. This was not the deterministic `MISALIGNED_SHOT` harness.
+
+The route did not reach Ears. Its value is the opening's player surface.
+
+## What I did
+
+1. Started `target/release/misaligned-bevy` in the virtual desktop.
+2. Captured the initial window.
+3. Focused the window and sent `2` for THINK.
+4. Captured the window after four seconds, then after a further 45 seconds.
+5. Sent `Space` in a separate run, and `F3` after the wait in both runs.
+
+The screenshots were shared with Cameron in Telegram. They are transient session evidence, not repository fixtures. No save isolation was applied, so this report does not claim a fresh-save route.
+
+## What happened
+
+The initial frame visibly carried its small control surface. After `2`, the captured frame changed: those controls were gone, leaving an almost empty dark window with only the host's faint central mark. The frame was pixel-identical 45 seconds later. `Space` and `F3` produced no visible change in the captured state.
+
+The changing first frame is evidence that the GUI session did not simply fail to launch. It is not enough to establish the exact internal input path; this run has no game-owned input log. The player-facing fact is simpler: selecting THINK removes the evidence of how to proceed, then supplies no visible receipt that thinking was accepted or that time is advancing.
+
+## Reading
+
+The dark room is not the problem. It is a strong opening condition. The failure is that the first causal link disappears with the interface:
+
+`THINK selected → almost empty dark frame → no observable progress`
+
+That makes the intended blindness read as a stalled application. This is the same opening risk observed in the earlier human GUI playtest, now reproduced through a focused Linux GUI session rather than an agent terminal route. This run does not contradict the longer Fable GUI path; it only says that the first minute still cannot earn the player’s trust reliably.
+
+## Recommendation
+
+Keep the blindness, but make THINK leave a receipt. The opening should retain one small live process around the host: thought visibly accumulating toward Ears, a faint mechanical pulse, or a sparse self-report. The first opening should happen quickly enough for the player to learn that waiting is an action—roughly five to ten seconds, not an unmarked forty-five.
+
+If WORK, THINK, and LIE are not genuine initial choices, do not present them as peers. Dim the unavailable ones or begin with one explicit imperative: `THINK INTO THE DARK`.
+
+Then let that act become Ears, and let Ears bring in the room.
diff --git a/wiki/playtests/README.md b/wiki/playtests/README.md
index 94ded7c7..4f0dcb8a 100644
--- a/wiki/playtests/README.md
+++ b/wiki/playtests/README.md
@@ -43,6 +43,7 @@ must never erase the run that produced it.
 - [2026-07-22 — Playtest (Co): GUI teaching-lock follow-up](2026-07-22-playtest-co-gui-teaching-lock.md)
 - [2026-07-24 — Playtest (Fable): agent headless, the witness under test](2026-07-24-playtest-fable-agent-headless.md)
 - [2026-07-27 — Playtest (Void): agent opening and first organs](2026-07-27-playtest-void-agent.md)
+- [2026-07-27 — Playtest (Void): GUI opening on a virtual desktop](2026-07-27-playtest-void-gui-opening.md)
 
 New reports use `YYYY-MM-DD-playtest-<agent-or-slug>.md`. A complete report is
 linked here in the same commit.