feat(voice): audio transcription for voice messages with Local Whisper support

console/src/api/modules/agent.ts

@@ -41,4 +41,12 @@ export const agentApi = {
       method: "PUT",
       body: JSON.stringify({ language }),
     }),
+
+  getAudioMode: () => request<{ audio_mode: string }>("/agent/audio-mode"),
+
+  updateAudioMode: (audio_mode: string) =>
+    request<{ audio_mode: string }>("/agent/audio-mode", {
+      method: "PUT",
+      body: JSON.stringify({ audio_mode }),
+    }),
 };

console/src/components/MarkdownCopy/MarkdownCopy.tsx

@@ -77,8 +77,8 @@ export function MarkdownCopy({
       localShowMarkdown && !(editable && !textareaProps.disabled)
         ? content
         : editable
-        ? editContent
-        : content;
+          ? editContent
+          : content;
 
     if (!contentToCopy) return;
 

console/src/layouts/MainLayout/index.tsx

@@ -19,6 +19,7 @@ import ModelsPage from "../../pages/Settings/Models";
 import EnvironmentsPage from "../../pages/Settings/Environments";
 import SecurityPage from "../../pages/Settings/Security";
 import TokenUsagePage from "../../pages/Settings/TokenUsage";
+import VoiceTranscriptionPage from "../../pages/Settings/VoiceTranscription";
 
 const { Content } = Layout;
 
@@ -38,6 +39,7 @@ const pathToKey: Record<string, string> = {
   "/agent-config": "agent-config",
   "/security": "security",
   "/token-usage": "token-usage",
+  "/voice-transcription": "voice-transcription",
 };
 
 export default function MainLayout() {
@@ -84,6 +86,10 @@ export default function MainLayout() {
                 <Route path="/agent-config" element={<AgentConfigPage />} />
                 <Route path="/security" element={<SecurityPage />} />
                 <Route path="/token-usage" element={<TokenUsagePage />} />
+                <Route
+                  path="/voice-transcription"
+                  element={<VoiceTranscriptionPage />}
+                />
               </Routes>
             )}
           </div>

console/src/layouts/Sidebar.tsx

@@ -36,6 +36,7 @@ import {
   Copy,
   Check,
   BarChart3,
+  Mic,
 } from "lucide-react";
 import api from "../api";
 import styles from "./index.module.less";
@@ -66,6 +67,7 @@ const KEY_TO_PATH: Record<string, string> = {
   "agent-config": "/agent-config",
   security: "/security",
   "token-usage": "/token-usage",
+  "voice-transcription": "/voice-transcription",
 };
 
 const UPDATE_MD: Record<string, string> = {
@@ -255,8 +257,8 @@ export default function Sidebar({ selectedKey }: SidebarProps) {
     const lang = i18n.language?.startsWith("zh")
       ? "zh"
       : i18n.language?.startsWith("ru")
-      ? "ru"
-      : "en";
+        ? "ru"
+        : "en";
     const faqLang = lang === "zh" ? "zh" : "en";
     const url = `https://copaw.agentscope.io/docs/faq.${faqLang}.md`;
     fetch(url, { cache: "no-cache" })
@@ -268,7 +270,7 @@ export default function Sidebar({ selectedKey }: SidebarProps) {
         setUpdateMarkdown(
           match && lang !== "ru"
             ? match[0].trim()
-            : UPDATE_MD[lang] ?? UPDATE_MD.en,
+            : (UPDATE_MD[lang] ?? UPDATE_MD.en),
         );
       })
       .catch(() => {
@@ -353,6 +355,11 @@ export default function Sidebar({ selectedKey }: SidebarProps) {
           label: t("nav.tokenUsage"),
           icon: <BarChart3 size={16} />,
         },
+        {
+          key: "voice-transcription",
+          label: t("nav.voiceTranscription"),
+          icon: <Mic size={16} />,
+        },
       ],
     },
   ];

console/src/locales/en.json

@@ -40,7 +40,23 @@
     "models": "Models",
     "environments": "Environments",
     "security": "Security",
-    "tokenUsage": "Token Usage"
+    "tokenUsage": "Token Usage",
+    "voiceTranscription": "Voice Transcription"
+  },
+  "voiceTranscription": {
+    "title": "Voice Transcription",
+    "description": "Configure how incoming audio and voice messages are handled.",
+    "loadFailed": "Failed to load audio mode settings",
+    "saveSuccess": "Audio mode saved",
+    "saveFailed": "Failed to save audio mode",
+    "audioModeLabel": "Audio Mode",
+    "audioModeDescription": "Choose how voice messages from channels (Discord, Telegram, etc.) are processed before being sent to the model.",
+    "modeAuto": "Auto",
+    "modeAutoDesc": "Try transcription first, fall back to native audio",
+    "modeTranscribe": "Transcribe",
+    "modeTranscribeDesc": "Always convert audio to text (works with all models)",
+    "modeNative": "Native",
+    "modeNativeDesc": "Send audio directly to the model (requires ffmpeg, only for audio-capable models)"
   },
   "workspace": {
     "title": "WorkSpace",

console/src/locales/zh.json

@@ -40,7 +40,8 @@
     "models": "模型",
     "environments": "环境变量",
     "security": "安全",
-    "tokenUsage": "Token 消耗"
+    "tokenUsage": "Token 消耗",
+    "voiceTranscription": "语音转写"
   },
   "workspace": {
     "title": "工作区",
@@ -644,5 +645,20 @@
         "TOOL_CMD_DANGEROUS_MV": "检测可能意外移动或覆盖文件的 mv 命令"
       }
     }
+  },
+  "voiceTranscription": {
+    "title": "语音转写",
+    "description": "配置如何处理收到的音频和语音消息。",
+    "loadFailed": "加载音频模式设置失败",
+    "saveSuccess": "音频模式已保存",
+    "saveFailed": "保存音频模式失败",
+    "audioModeLabel": "音频模式",
+    "audioModeDescription": "选择来自频道（Discord、Telegram 等）的语音消息在发送给模型之前如何处理。",
+    "modeAuto": "自动",
+    "modeAutoDesc": "优先转写，失败时回退到原生音频",
+    "modeTranscribe": "转写",
+    "modeTranscribeDesc": "始终将音频转为文字（适用于所有模型）",
+    "modeNative": "原生",
+    "modeNativeDesc": "直接发送音频给模型（需要 ffmpeg，仅支持音频的模型可用）"
   }
 }

console/src/pages/Control/Sessions/index.tsx

@@ -56,9 +56,8 @@ function SessionsPage() {
     let filtered: Session[] = sessions;
 
     if (filterUserId) {
-      filtered = filtered.filter(
-        (session: Session) =>
-          session.user_id?.toLowerCase().includes(filterUserId.toLowerCase()),
+      filtered = filtered.filter((session: Session) =>
+        session.user_id?.toLowerCase().includes(filterUserId.toLowerCase()),
       );
     }
 

console/src/pages/Settings/Models/components/cards/RemoteProviderCard.tsx

@@ -87,23 +87,23 @@ export function RemoteProviderCard({
   const statusLabel = isAvailable
     ? t("models.providerAvailable")
     : isConfigured
-    ? t("models.providerNoModels")
-    : t("models.providerNotConfigured");
+      ? t("models.providerNoModels")
+      : t("models.providerNotConfigured");
   const statusType = isAvailable
     ? "enabled"
     : isConfigured
-    ? "partial"
-    : "disabled";
+      ? "partial"
+      : "disabled";
   const statusDotColor = isAvailable
     ? "#52c41a"
     : isConfigured
-    ? "#faad14"
-    : "#d9d9d9";
+      ? "#faad14"
+      : "#d9d9d9";
   const statusDotShadow = isAvailable
     ? "0 0 0 2px rgba(82, 196, 26, 0.2)"
     : isConfigured
-    ? "0 0 0 2px rgba(250, 173, 20, 0.2)"
-    : "none";
+      ? "0 0 0 2px rgba(250, 173, 20, 0.2)"
+      : "none";
 
   return (
     <Card
@@ -135,8 +135,8 @@ export function RemoteProviderCard({
                 statusType === "enabled"
                   ? styles.enabled
                   : statusType === "partial"
-                  ? styles.partial
-                  : styles.disabled
+                    ? styles.partial
+                    : styles.disabled
               }`}
             >
               {statusLabel}

console/src/pages/Settings/Models/components/modals/ProviderConfigModal.tsx

@@ -14,8 +14,10 @@ import api from "../../../../../api";
 import { useTranslation } from "react-i18next";
 import styles from "../../index.module.less";
 
-interface ProviderConfigFormValues
-  extends Omit<ProviderConfigRequest, "generate_kwargs"> {
+interface ProviderConfigFormValues extends Omit<
+  ProviderConfigRequest,
+  "generate_kwargs"
+> {
   generate_kwargs_text?: string;
 }
 

console/src/pages/Settings/VoiceTranscription/index.module.less

@@ -0,0 +1,73 @@
+.page {
+  padding: 24px;
+  height: 100%;
+  overflow-y: auto;
+  width: 100%;
+}
+
+.header {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+  margin-bottom: 24px;
+}
+
+.title {
+  font-size: 24px;
+  font-weight: 600;
+  margin: 0 0 8px 0;
+  color: #000;
+}
+
+.description {
+  font-size: 14px;
+  color: #666;
+  margin: 0;
+}
+
+.card {
+  width: 100%;
+  margin-bottom: 16px;
+
+  &:hover {
+    border: 1px solid #615ced;
+    box-shadow: 0 12px 32px rgba(0, 0, 0, 0.08);
+  }
+}
+
+.cardTitle {
+  font-size: 16px;
+  font-weight: 600;
+  margin: 0 0 4px 0;
+}
+
+.cardDescription {
+  font-size: 13px;
+  color: #888;
+  margin: 0 0 16px 0;
+}
+
+.optionLabel {
+  font-weight: 500;
+  margin-right: 8px;
+}
+
+.optionDescription {
+  font-size: 13px;
+  color: #888;
+}
+
+.centerState {
+  display: flex;
+  flex-direction: column;
+  align-items: center;
+  justify-content: center;
+  height: 400px;
+}
+
+.footerActions {
+  display: flex;
+  justify-content: flex-end;
+  align-items: center;
+  padding: 16px 24px;
+}