diff --git a/src/vs/workbench/contrib/agentsVoice/browser/agentsVoice.contribution.ts b/src/vs/workbench/contrib/agentsVoice/browser/agentsVoice.contribution.ts
index 7644f12f89e88b..2bf118c4fc5cf7 100644
--- a/src/vs/workbench/contrib/agentsVoice/browser/agentsVoice.contribution.ts
+++ b/src/vs/workbench/contrib/agentsVoice/browser/agentsVoice.contribution.ts
@@ -37,7 +37,6 @@ import { IStorageService, StorageScope, StorageTarget } from '../../../../platfo
 
 import { AgentsVoiceStorageKeys, AGENTS_VOICE_CONNECTED, AGENTS_VOICE_CONNECTING, AGENTS_VOICE_LISTENING } from '../common/agentsVoice.js';
 import { IConfigurationService } from '../../../../platform/configuration/common/configuration.js';
-import { IQuickInputService } from '../../../../platform/quickinput/common/quickInput.js';
 import { ITelemetryService } from '../../../../platform/telemetry/common/telemetry.js';
 import { INotificationService } from '../../../../platform/notification/common/notification.js';
 import {
@@ -537,77 +536,9 @@ registerAction2(class extends Action2 {
 	}
 });
 
-// --- Select Microphone Command ---
-
-registerAction2(class extends Action2 {
-	constructor() {
-		super({
-			id: 'agentsVoice.selectMicrophone',
-			title: nls.localize2('agentsVoice.selectMicrophone', "Voice: Select Microphone"),
-			f1: true,
-			precondition: ContextKeyExpr.equals('config.agents.voice.enabled', true),
-		});
-	}
-	async run(accessor: ServicesAccessor): Promise<void> {
-		const quickInputService = accessor.get(IQuickInputService);
-		const storageService = accessor.get(IStorageService);
-
-		const devices = await navigator.mediaDevices.enumerateDevices();
-
-		// Filter out the virtual "default"/"communications" entries (which duplicate a real
-		// device) and de-duplicate by deviceId so a single microphone shows up only once.
-		const seenDeviceIds = new Set<string>();
-		const audioInputs = devices.filter(d => {
-			if (d.kind !== 'audioinput' || d.deviceId === 'default' || d.deviceId === 'communications') {
-				return false;
-			}
-			if (seenDeviceIds.has(d.deviceId)) {
-				return false;
-			}
-			seenDeviceIds.add(d.deviceId);
-			return true;
-		});
-
-		if (audioInputs.length === 0) {
-			quickInputService.pick([{ label: nls.localize('noMicrophones', "No microphones found") }]);
-			return;
-		}
-
-		const currentDeviceId = storageService.get(AgentsVoiceStorageKeys.MicrophoneDevice, StorageScope.APPLICATION, '');
-
-		type DevicePickItem = { label: string; description?: string; deviceId: string };
-		const items: DevicePickItem[] = [];
-
-		// "System Default" entry — clears the stored device so the OS default is always used
-		items.push({
-			label: nls.localize('systemDefault', "System Default"),
-			description: currentDeviceId === '' ? nls.localize('current', "(current)") : undefined,
-			deviceId: '',
-		});
-
-		for (const d of audioInputs) {
-			const label = d.label || nls.localize('unknownDevice', "Unknown Device ({0})", d.deviceId.slice(0, 8));
-			items.push({
-				label,
-				description: d.deviceId === currentDeviceId ? nls.localize('current', "(current)") : undefined,
-				deviceId: d.deviceId,
-			});
-		}
-
-		const picked = await quickInputService.pick(items, {
-			placeHolder: nls.localize('selectMic', "Select a microphone for Voice Mode"),
-		});
-
-		if (picked) {
-			const selection = picked as DevicePickItem;
-			if (selection.deviceId) {
-				storageService.store(AgentsVoiceStorageKeys.MicrophoneDevice, selection.deviceId, StorageScope.APPLICATION, StorageTarget.MACHINE);
-			} else {
-				storageService.remove(AgentsVoiceStorageKeys.MicrophoneDevice, StorageScope.APPLICATION);
-			}
-		}
-	}
-});
+// Microphone selection is shared with dictation via the single
+// `workbench.action.chat.selectSpeechToTextMicrophone` command (see
+// chatSpeechToTextActions.ts), so Voice Mode no longer registers its own.
 
 // --- Settings ---
 
diff --git a/src/vs/workbench/contrib/chat/browser/actions/chatSpeechToTextActions.ts b/src/vs/workbench/contrib/chat/browser/actions/chatSpeechToTextActions.ts
index f47b07b2d68e1a..0ecfb0d7a61a05 100644
--- a/src/vs/workbench/contrib/chat/browser/actions/chatSpeechToTextActions.ts
+++ b/src/vs/workbench/contrib/chat/browser/actions/chatSpeechToTextActions.ts
@@ -32,7 +32,7 @@ import { cancelDictation, isDictating, startDictation, stopDictation } from '../
 
 // Gate on `ChatContextKeys.enabled` so the dictation UI and its commands are
 // hidden (not just disabled) when the user has turned AI features off; without
-// it the F1 "Dictate: Select Microphone" command stays discoverable.
+// it the F1 "Select Microphone" command stays discoverable.
 export const ChatSpeechToTextConfigured = ContextKeyExpr.and(ChatContextKeys.enabled, ContextKeyExpr.has(ChatContextKeys.speechToTextConfigured.key));
 /** True while the selected dictation backend is preparing. */
 export const ChatSpeechToTextPreparing = ContextKeyExpr.has(ChatContextKeys.speechToTextPreparing.key);
@@ -292,10 +292,12 @@ class SelectSpeechToTextMicrophoneAction extends Action2 {
 	constructor() {
 		super({
 			id: SelectSpeechToTextMicrophoneAction.ID,
-			title: localize2('chat.speechToText.selectMicrophone', "Dictate: Select Microphone"),
+			title: localize2('chat.speechToText.selectMicrophone', "Select Microphone"),
 			category: CHAT_CATEGORY,
 			f1: true,
-			precondition: ChatSpeechToTextConfigured,
+			// Shared by dictation and Voice Mode (both persist to the same
+			// device), so stay available whenever either feature is enabled.
+			precondition: ContextKeyExpr.or(ChatSpeechToTextConfigured, ContextKeyExpr.equals('config.agents.voice.enabled', true)),
 		});
 	}