diff --git a/docs/cli/settings.md b/docs/cli/settings.md
index a5c7ecae87..d39a0e18f7 100644
--- a/docs/cli/settings.md
+++ b/docs/cli/settings.md
@@ -166,7 +166,7 @@ they appear in the UI.
| Gemma Models | `experimental.gemma` | Enable access to Gemma 4 models via Gemini API. | `true` |
| Voice Mode | `experimental.voiceMode` | Enable experimental voice dictation and commands (/voice, /voice model). | `false` |
| Voice Activation Mode | `experimental.voice.activationMode` | How to trigger voice recording with the Space key. | `"push-to-talk"` |
-| Voice Transcription Backend | `experimental.voice.backend` | The backend to use for voice transcription. | `"gemini-live"` |
+| Voice Transcription Backend | `experimental.voice.backend` | The backend to use for voice transcription. Note: When using the Gemini Live backend, voice recordings are sent to Google Cloud for transcription. | `"gemini-live"` |
| Whisper Model | `experimental.voice.whisperModel` | The Whisper model to use for local transcription. | `"ggml-base.en.bin"` |
| Voice Stop Grace Period (ms) | `experimental.voice.stopGracePeriodMs` | How long to wait for final transcription after stopping recording. | `1000` |
| Enable Git Worktrees | `experimental.worktrees` | Enable automated Git worktree management for parallel work. | `false` |
diff --git a/docs/reference/configuration.md b/docs/reference/configuration.md
index f0eaafc27c..3498634dd1 100644
--- a/docs/reference/configuration.md
+++ b/docs/reference/configuration.md
@@ -1774,7 +1774,9 @@ their corresponding top-level category object in your `settings.json` file.
- **Values:** `"push-to-talk"`, `"toggle"`
- **`experimental.voice.backend`** (enum):
- - **Description:** The backend to use for voice transcription.
+ - **Description:** The backend to use for voice transcription. Note: When
+ using the Gemini Live backend, voice recordings are sent to Google Cloud for
+ transcription.
- **Default:** `"gemini-live"`
- **Values:** `"gemini-live"`, `"whisper"`
diff --git a/packages/cli/src/config/settingsSchema.ts b/packages/cli/src/config/settingsSchema.ts
index 225d3d8ac0..fa941c9a01 100644
--- a/packages/cli/src/config/settingsSchema.ts
+++ b/packages/cli/src/config/settingsSchema.ts
@@ -2099,7 +2099,11 @@ const SETTINGS_SCHEMA = {
category: 'Experimental',
requiresRestart: false,
default: 'gemini-live',
- description: 'The backend to use for voice transcription.',
+ description: oneLine`
+ The backend to use for voice transcription. Note: When using the
+ Gemini Live backend, voice recordings are sent to Google Cloud for
+ transcription.
+ `,
showInDialog: true,
options: [
{ value: 'gemini-live', label: 'Gemini Live API (Cloud)' },
diff --git a/packages/cli/src/ui/components/VoiceModelDialog.test.tsx b/packages/cli/src/ui/components/VoiceModelDialog.test.tsx
new file mode 100644
index 0000000000..7ec081b032
--- /dev/null
+++ b/packages/cli/src/ui/components/VoiceModelDialog.test.tsx
@@ -0,0 +1,92 @@
+/**
+ * @license
+ * Copyright 2026 Google LLC
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import { describe, it, expect, vi } from 'vitest';
+import { renderWithProviders } from '../../test-utils/render.js';
+import { createMockSettings } from '../../test-utils/settings.js';
+import { VoiceModelDialog } from './VoiceModelDialog.js';
+import { act } from 'react';
+import { waitFor } from '../../test-utils/async.js';
+import { SettingScope } from '../../config/settings.js';
+
+vi.mock('@google/gemini-cli-core', async () => {
+ const actual = await vi.importActual('@google/gemini-cli-core');
+ return {
+ ...actual,
+ isBinaryAvailable: vi.fn().mockReturnValue(true),
+ WhisperModelManager: vi.fn().mockImplementation(() => ({
+ isModelInstalled: vi.fn().mockReturnValue(false),
+ on: vi.fn(),
+ off: vi.fn(),
+ downloadModel: vi.fn(),
+ })),
+ };
+});
+
+describe('VoiceModelDialog', () => {
+ it('should display a privacy warning when Gemini Live API (Cloud) is selected', async () => {
+ const onClose = vi.fn();
+ const { lastFrame, waitUntilReady } = await renderWithProviders(
+ ,
+ );
+
+ await waitUntilReady();
+
+ const frame = lastFrame();
+ expect(frame).toContain('Gemini Live API (Cloud)');
+ expect(frame).toContain('When using the Gemini Live backend');
+ });
+
+ it('should NOT display a privacy warning when Whisper (Local) is highlighted', async () => {
+ const onClose = vi.fn();
+ const { lastFrame, waitUntilReady, stdin } = await renderWithProviders(
+ ,
+ );
+
+ await waitUntilReady();
+
+ // Verify warning is present for default (Gemini Live)
+ expect(lastFrame()).toContain('When using the Gemini Live backend');
+
+ // Arrow Down to highlight Whisper
+ await act(async () => {
+ stdin.write('\u001b[B');
+ });
+
+ await waitFor(() => {
+ const frame = lastFrame();
+ expect(frame).toContain('Whisper (Local)');
+ expect(frame).not.toContain('When using the Gemini Live backend');
+ });
+ });
+
+ it('should update settings and close dialog when a backend is selected', async () => {
+ const onClose = vi.fn();
+ const settings = createMockSettings();
+ const setValueSpy = vi.spyOn(settings, 'setValue');
+
+ const { waitUntilReady, stdin } = await renderWithProviders(
+ ,
+ { settings },
+ );
+
+ await waitUntilReady();
+
+ // Select Gemini Live (it's already highlighted, just press Enter)
+ await act(async () => {
+ stdin.write('\r');
+ });
+
+ await waitFor(() => {
+ expect(setValueSpy).toHaveBeenCalledWith(
+ SettingScope.User,
+ 'experimental.voice.backend',
+ 'gemini-live',
+ );
+ expect(onClose).toHaveBeenCalled();
+ });
+ });
+});
diff --git a/packages/cli/src/ui/components/VoiceModelDialog.tsx b/packages/cli/src/ui/components/VoiceModelDialog.tsx
index f340a5ccf4..e882c89235 100644
--- a/packages/cli/src/ui/components/VoiceModelDialog.tsx
+++ b/packages/cli/src/ui/components/VoiceModelDialog.tsx
@@ -18,6 +18,7 @@ import {
type WhisperModelProgress,
} from '@google/gemini-cli-core';
import { CliSpinner } from './CliSpinner.js';
+import { WarningMessage } from './messages/WarningMessage.js';
interface VoiceModelDialogProps {
onClose: () => void;
@@ -68,6 +69,9 @@ export function VoiceModelDialog({
const currentWhisperModel =
settings.merged.experimental.voice?.whisperModel ?? 'ggml-base.en.bin';
+ const [highlightedBackend, setHighlightedBackend] =
+ useState(currentBackend);
+
const handleKeypress = useCallback(
(key: Key) => {
if (key.name === 'escape') {
@@ -101,6 +105,10 @@ export function VoiceModelDialog({
[setSetting, onClose],
);
+ const handleBackendHighlight = useCallback((value: string) => {
+ setHighlightedBackend(value);
+ }, []);
+
const handleWhisperModelSelect = useCallback(
async (modelName: string) => {
if (modelManager.isModelInstalled(modelName)) {
@@ -203,14 +211,22 @@ export function VoiceModelDialog({
) : (
-
+
{view === 'backend' ? (
-
+ <>
+
+ {highlightedBackend === 'gemini-live' && (
+
+
+
+ )}
+
) : (