From ad6bb70c2013903ae8ea6bc7c9eb641f846b3da6 Mon Sep 17 00:00:00 2001
From: Yun Chan <yunchan@twentyoz.kr>
Date: Thu, 24 Sep 2026 22:08:05 +0900
Subject: [PATCH] fix(settings): show the live caption model under the
 dictation model

The caption model selector now sits directly under the dictation model in
the STT tab's Local Whisper card, where people look for it, and stays in the
General tab's caption section as well; both use one shared component. The
first selector is renamed to 'Dictation model' so the two are told apart.
---
 .../components/CaptionModelSelect.tsx         | 41 +++++++++++++++++++
 .../src/renderer/components/STTTab.tsx        |  9 ++++
 .../src/renderer/components/SettingsModal.tsx | 29 +++----------
 packages/i18n/src/locales/de.json             |  2 +-
 packages/i18n/src/locales/en.json             |  2 +-
 packages/i18n/src/locales/es.json             |  2 +-
 packages/i18n/src/locales/fr.json             |  2 +-
 packages/i18n/src/locales/ja.json             |  2 +-
 packages/i18n/src/locales/ko.json             |  2 +-
 packages/i18n/src/locales/pt.json             |  2 +-
 packages/i18n/src/locales/ru.json             |  2 +-
 packages/i18n/src/locales/th.json             |  2 +-
 packages/i18n/src/locales/vi.json             |  2 +-
 packages/i18n/src/locales/zh-TW.json          |  2 +-
 packages/i18n/src/locales/zh.json             |  2 +-
 15 files changed, 68 insertions(+), 35 deletions(-)
 create mode 100644 apps/desktop/src/renderer/components/CaptionModelSelect.tsx

diff --git a/apps/desktop/src/renderer/components/CaptionModelSelect.tsx b/apps/desktop/src/renderer/components/CaptionModelSelect.tsx
new file mode 100644
index 0000000..a02163e
--- /dev/null
+++ b/apps/desktop/src/renderer/components/CaptionModelSelect.tsx
@@ -0,0 +1,41 @@
+// 실시간 자막 전용 로컬 Whisper 모델 선택 — STT 탭과 일반 탭 자막 섹션이 함께 쓴다.
+
+import { FormControl, InputLabel, MenuItem, Select, Typography } from '@mui/material'
+import { d3roPalette, d3roTypo } from '@d3ro/ui/theme'
+import { useI18n } from '@d3ro/i18n'
+import type { STTModel } from '@d3ro/core/types'
+
+/** "받아쓰기와 같게" — 설정에는 null 로 저장한다 */
+const SAME_AS_DICTATION = '__same__'
+
+interface CaptionModelSelectProps {
+  value: string | null | undefined
+  /** 선택지로 보일 모델 — 받아 둔 모델만 넘긴다 */
+  models: readonly STTModel[]
+  onChange: (modelId: string | null) => void
+  fullWidth?: boolean
+}
+
+export function CaptionModelSelect({ value, models, onChange, fullWidth }: CaptionModelSelectProps): React.ReactElement {
+  const { t } = useI18n()
+  return (
+    <FormControl size="small" fullWidth={fullWidth}>
+      <InputLabel>{t('settings.captionModel')}</InputLabel>
+      <Select
+        label={t('settings.captionModel')}
+        value={value ?? SAME_AS_DICTATION}
+        onChange={(e) => onChange(e.target.value === SAME_AS_DICTATION ? null : e.target.value)}
+      >
+        <MenuItem value={SAME_AS_DICTATION}>{t('settings.captionModel.same')}</MenuItem>
+        {models.map((m) => (
+          <MenuItem key={m.id} value={m.id}>
+            {m.name}
+          </MenuItem>
+        ))}
+      </Select>
+      <Typography sx={{ mt: 0.5, fontSize: d3roTypo.meta.size, color: d3roPalette.text.secondary }}>
+        {t('settings.captionModel.desc')}
+      </Typography>
+    </FormControl>
+  )
+}
diff --git a/apps/desktop/src/renderer/components/STTTab.tsx b/apps/desktop/src/renderer/components/STTTab.tsx
index 3c2f53e..a6e2415 100644
--- a/apps/desktop/src/renderer/components/STTTab.tsx
+++ b/apps/desktop/src/renderer/components/STTTab.tsx
@@ -50,6 +50,7 @@ import type {
   STTModel,
 } from '@d3ro/core/types'
 import { CodexOAuthGuideModal } from './CodexOAuthGuideModal'
+import { CaptionModelSelect } from './CaptionModelSelect'
 
 interface STTTabProps {
   config: Partial<AppConfig>
@@ -362,6 +363,14 @@ export function STTTab({ config, updateConfig }: STTTabProps): React.ReactElemen
               </Select>
             </FormControl>
 
+            {/* 실시간 자막 전용 모델 — 1초마다 다시 인식하므로 받아쓰기와 다른 모델이 나을 수 있다 */}
+            <CaptionModelSelect
+              fullWidth
+              value={config.captionSttModelId}
+              models={localModels.filter((m) => m.downloaded)}
+              onChange={(modelId) => updateConfig('captionSttModelId', modelId)}
+            />
+
             {/* 모델 다운로드 진행 바 또는 다운로드 버튼 */}
             {(() => {
               const selected = localModels.find((m) => m.id === (config.sttModelId ?? 'large-v3-turbo'))
diff --git a/apps/desktop/src/renderer/components/SettingsModal.tsx b/apps/desktop/src/renderer/components/SettingsModal.tsx
index 626a84c..61495c1 100644
--- a/apps/desktop/src/renderer/components/SettingsModal.tsx
+++ b/apps/desktop/src/renderer/components/SettingsModal.tsx
@@ -28,6 +28,7 @@ import { Led } from '@d3ro/ui/components/ds'
 import { LicenseTab } from './LicenseTab'
 import { CloudSyncSection } from './CloudSyncSection'
 import { STTTab } from './STTTab'
+import { CaptionModelSelect } from './CaptionModelSelect'
 import { InputConsentPanel } from './input-insights/InputConsentPanel'
 import { KeyBindingField } from './keybinding/KeyBindingField'
 import { asTranslationKey } from './keybinding/translation-key'
@@ -64,9 +65,6 @@ const ACTION_GROUP_LABEL_KEYS: Readonly<Record<KeyBindingActionGroup, string>> =
 
 const ACTION_GROUP_ORDER: readonly KeyBindingActionGroup[] = ['voice', 'window', 'input']
 
-/** 자막 모델 선택지 "받아쓰기와 같게" — 설정에는 null 로 저장한다 */
-const SAME_AS_DICTATION = '__same__'
-
 export function SettingsModal({ open, initialTab = 0, onClose }: SettingsModalProps): React.ReactElement {
   const { t, locale, setLocale } = useI18n()
   const [activeTab, setActiveTab] = useState(initialTab)
@@ -493,26 +491,11 @@ export function SettingsModal({ open, initialTab = 0, onClose }: SettingsModalPr
                 </Select>
               </FormControl>
 
-              <FormControl size="small">
-                <InputLabel>{t('settings.captionModel')}</InputLabel>
-                <Select
-                  label={t('settings.captionModel')}
-                  value={config.captionSttModelId ?? SAME_AS_DICTATION}
-                  onChange={(e) =>
-                    updateConfig('captionSttModelId', e.target.value === SAME_AS_DICTATION ? null : e.target.value)
-                  }
-                >
-                  <MenuItem value={SAME_AS_DICTATION}>{t('settings.captionModel.same')}</MenuItem>
-                  {downloadedSttModels.map((m) => (
-                    <MenuItem key={m.id} value={m.id}>
-                      {m.name}
-                    </MenuItem>
-                  ))}
-                </Select>
-                <Typography sx={{ mt: 0.5, fontSize: d3roTypo.meta.size, color: d3roPalette.text.secondary }}>
-                  {t('settings.captionModel.desc')}
-                </Typography>
-              </FormControl>
+              <CaptionModelSelect
+                value={config.captionSttModelId}
+                models={downloadedSttModels}
+                onChange={(modelId) => updateConfig('captionSttModelId', modelId)}
+              />
 
               <FormControlLabel
                 sx={{ mr: 0, alignItems: 'flex-start' }}
diff --git a/packages/i18n/src/locales/de.json b/packages/i18n/src/locales/de.json
index ea615ea..0a51d8e 100644
--- a/packages/i18n/src/locales/de.json
+++ b/packages/i18n/src/locales/de.json
@@ -99,7 +99,7 @@
   "settings.insertMethod": "Einfügemethode",
   "settings.insertClipboard": "Zwischenablage (Ctrl+V)",
   "settings.insertKeyboard": "Tastatureingabe",
-  "settings.whisperModel": "Whisper-Modell",
+  "settings.whisperModel": "Diktiermodell",
   "settings.model.tiny": "tiny (39 MB, am schnellsten)",
   "settings.model.base": "base (74 MB, ausgewogen)",
   "settings.model.small": "small (244 MB, gut)",
diff --git a/packages/i18n/src/locales/en.json b/packages/i18n/src/locales/en.json
index 830c4b3..03c3b7e 100644
--- a/packages/i18n/src/locales/en.json
+++ b/packages/i18n/src/locales/en.json
@@ -139,7 +139,7 @@
   "settings.sttProvider.custom": "Custom (OpenAI-Compatible)",
   "settings.sttFallbackToLocal": "Auto Fallback to Local Whisper",
   "settings.sttFallbackToLocalHint": "Automatically switch to offline local Whisper if cloud STT fails",
-  "settings.whisperModel": "Whisper Model",
+  "settings.whisperModel": "Dictation model",
   "settings.model.tiny": "tiny (39 MB, fastest)",
   "settings.model.base": "base (74 MB, balanced)",
   "settings.model.small": "small (244 MB, good)",
diff --git a/packages/i18n/src/locales/es.json b/packages/i18n/src/locales/es.json
index b1211c9..cc884b0 100644
--- a/packages/i18n/src/locales/es.json
+++ b/packages/i18n/src/locales/es.json
@@ -99,7 +99,7 @@
   "settings.insertMethod": "Método de inserción",
   "settings.insertClipboard": "Portapapeles (Ctrl+V)",
   "settings.insertKeyboard": "Escritura con teclado",
-  "settings.whisperModel": "Modelo Whisper",
+  "settings.whisperModel": "Modelo de dictado",
   "settings.model.tiny": "tiny (39 MB, más rápido)",
   "settings.model.base": "base (74 MB, equilibrado)",
   "settings.model.small": "small (244 MB, bueno)",
diff --git a/packages/i18n/src/locales/fr.json b/packages/i18n/src/locales/fr.json
index 61648f6..3a5c8d3 100644
--- a/packages/i18n/src/locales/fr.json
+++ b/packages/i18n/src/locales/fr.json
@@ -99,7 +99,7 @@
   "settings.insertMethod": "Méthode d'insertion",
   "settings.insertClipboard": "Presse-papiers (Ctrl+V)",
   "settings.insertKeyboard": "Frappe au clavier",
-  "settings.whisperModel": "Modèle Whisper",
+  "settings.whisperModel": "Modèle de dictée",
   "settings.model.tiny": "tiny (39 Mo, le plus rapide)",
   "settings.model.base": "base (74 Mo, équilibré)",
   "settings.model.small": "small (244 Mo, bon)",
diff --git a/packages/i18n/src/locales/ja.json b/packages/i18n/src/locales/ja.json
index 5ace182..ab56694 100644
--- a/packages/i18n/src/locales/ja.json
+++ b/packages/i18n/src/locales/ja.json
@@ -99,7 +99,7 @@
   "settings.insertMethod": "挿入方式",
   "settings.insertClipboard": "クリップボード (Ctrl+V)",
   "settings.insertKeyboard": "キーボード入力",
-  "settings.whisperModel": "Whisperモデル",
+  "settings.whisperModel": "音声入力モデル",
   "settings.model.tiny": "tiny (39 MB, 最速)",
   "settings.model.base": "base (74 MB, バランス)",
   "settings.model.small": "small (244 MB, 良好)",
diff --git a/packages/i18n/src/locales/ko.json b/packages/i18n/src/locales/ko.json
index 288b46d..cd2c7e4 100644
--- a/packages/i18n/src/locales/ko.json
+++ b/packages/i18n/src/locales/ko.json
@@ -147,7 +147,7 @@
   "settings.sttProvider.custom": "커스텀 OpenAI 호환 엔드포인트",
   "settings.sttFallbackToLocal": "로컬 Whisper 자동 폴백",
   "settings.sttFallbackToLocalHint": "클라우드 STT 실패 시 오프라인 로컬 Whisper로 자동 전환",
-  "settings.whisperModel": "Whisper 모델",
+  "settings.whisperModel": "받아쓰기 모델",
   "settings.model.tiny": "tiny (39 MB, 가장 빠름)",
   "settings.model.base": "base (74 MB, 균형)",
   "settings.model.small": "small (244 MB, 양호)",
diff --git a/packages/i18n/src/locales/pt.json b/packages/i18n/src/locales/pt.json
index ae967f2..f7651b8 100644
--- a/packages/i18n/src/locales/pt.json
+++ b/packages/i18n/src/locales/pt.json
@@ -99,7 +99,7 @@
   "settings.insertMethod": "Método de inserção",
   "settings.insertClipboard": "Área de transferência (Ctrl+V)",
   "settings.insertKeyboard": "Digitação pelo teclado",
-  "settings.whisperModel": "Modelo Whisper",
+  "settings.whisperModel": "Modelo de ditado",
   "settings.model.tiny": "tiny (39 MB, mais rápido)",
   "settings.model.base": "base (74 MB, equilibrado)",
   "settings.model.small": "small (244 MB, bom)",
diff --git a/packages/i18n/src/locales/ru.json b/packages/i18n/src/locales/ru.json
index 6df5592..6cf3cb0 100644
--- a/packages/i18n/src/locales/ru.json
+++ b/packages/i18n/src/locales/ru.json
@@ -99,7 +99,7 @@
   "settings.insertMethod": "Метод вставки",
   "settings.insertClipboard": "Буфер обмена (Ctrl+V)",
   "settings.insertKeyboard": "Эмуляция клавиатуры",
-  "settings.whisperModel": "Модель Whisper",
+  "settings.whisperModel": "Модель диктовки",
   "settings.model.tiny": "tiny (39 МБ, самая быстрая)",
   "settings.model.base": "base (74 МБ, баланс)",
   "settings.model.small": "small (244 МБ, хорошая)",
diff --git a/packages/i18n/src/locales/th.json b/packages/i18n/src/locales/th.json
index 0cad497..fb6f61c 100644
--- a/packages/i18n/src/locales/th.json
+++ b/packages/i18n/src/locales/th.json
@@ -99,7 +99,7 @@
   "settings.insertMethod": "วิธีการแทรก",
   "settings.insertClipboard": "คลิปบอร์ด (Ctrl+V)",
   "settings.insertKeyboard": "จำลองการพิมพ์",
-  "settings.whisperModel": "โมเดล Whisper",
+  "settings.whisperModel": "โมเดลพิมพ์ด้วยเสียง",
   "settings.model.tiny": "tiny (39 MB, เร็วที่สุด)",
   "settings.model.base": "base (74 MB, สมดุล)",
   "settings.model.small": "small (244 MB, ดี)",
diff --git a/packages/i18n/src/locales/vi.json b/packages/i18n/src/locales/vi.json
index b993ca7..b1e4bc9 100644
--- a/packages/i18n/src/locales/vi.json
+++ b/packages/i18n/src/locales/vi.json
@@ -99,7 +99,7 @@
   "settings.insertMethod": "Phương thức chèn",
   "settings.insertClipboard": "Bộ nhớ tạm (Ctrl+V)",
   "settings.insertKeyboard": "Giả lập bàn phím",
-  "settings.whisperModel": "Mô hình Whisper",
+  "settings.whisperModel": "Mô hình đọc chính tả",
   "settings.model.tiny": "tiny (39 MB, nhanh nhất)",
   "settings.model.base": "base (74 MB, cân bằng)",
   "settings.model.small": "small (244 MB, tốt)",
diff --git a/packages/i18n/src/locales/zh-TW.json b/packages/i18n/src/locales/zh-TW.json
index 446cb9f..37dce4f 100644
--- a/packages/i18n/src/locales/zh-TW.json
+++ b/packages/i18n/src/locales/zh-TW.json
@@ -99,7 +99,7 @@
   "settings.insertMethod": "插入方式",
   "settings.insertClipboard": "剪貼簿（Ctrl+V）",
   "settings.insertKeyboard": "鍵盤輸入",
-  "settings.whisperModel": "Whisper模型",
+  "settings.whisperModel": "聽寫模型",
   "settings.model.tiny": "tiny（39 MB，最快）",
   "settings.model.base": "base（74 MB，均衡）",
   "settings.model.small": "small（244 MB，良好）",
diff --git a/packages/i18n/src/locales/zh.json b/packages/i18n/src/locales/zh.json
index 7acb49d..262f4eb 100644
--- a/packages/i18n/src/locales/zh.json
+++ b/packages/i18n/src/locales/zh.json
@@ -99,7 +99,7 @@
   "settings.insertMethod": "插入方式",
   "settings.insertClipboard": "剪贴板（Ctrl+V）",
   "settings.insertKeyboard": "键盘输入",
-  "settings.whisperModel": "Whisper模型",
+  "settings.whisperModel": "听写模型",
   "settings.model.tiny": "tiny（39 MB，最快）",
   "settings.model.base": "base（74 MB，均衡）",
   "settings.model.small": "small（244 MB，良好）",