diff --git a/webapp/components/device.tsx b/webapp/components/device.tsx
index 7659fee..8d80ffa 100644
--- a/webapp/components/device.tsx
+++ b/webapp/components/device.tsx
@@ -7,6 +7,7 @@ import {
 } from '../lib/device'
 
 import Loading from './svg/loading'
+import SvgSpeaker from './svg/speaker'
 import SvgAudio from './svg/audio'
 import SvgVideo from './svg/video'
 import { SvgPresentCancel, SvgPresentToAll } from './svg/present'
@@ -24,20 +25,25 @@ export default function DeviceBar(props: { streamId: string }) {
   const [permissionAudio, setPermissionAudio] = useState('')
   const [permissionVideo, setPermissionVideo] = useState('')
 
+  const [loadingSpeaker, setLoadingSpeaker] = useState(false)
   const [loadingAudio, setLoadingAudio] = useState(false)
   const [loadingVideo, setLoadingVideo] = useState(false)
   const [loadingScreen, setLoadingScreen] = useState(false)
 
   const {
     userStatus,
+    currentDeviceSpeaker,
     currentDeviceAudio,
     currentDeviceVideo,
+    setCurrentDeviceSpeaker,
     setCurrentDeviceAudio,
     setCurrentDeviceVideo,
+    toggleEnableSpeaker,
     toggleEnableAudio,
     toggleEnableVideo,
   } = useWhipClient(props.streamId)
 
+  const [deviceSpeaker, setDeviceSpeaker] = useState<Device[]>([deviceNone])
   const [deviceAudio, setDeviceAudio] = useState<Device[]>([deviceNone])
   const [deviceVideo, setDeviceVideo] = useState<Device[]>([deviceNone])
 
@@ -82,9 +88,15 @@ export default function DeviceBar(props: { streamId: string }) {
     
     const devices = (await navigator.mediaDevices.enumerateDevices()).filter(i => !!i.deviceId)
 
+    const speakers = devices.filter(i => i.kind === 'audiooutput').map(toDevice)
     const audios = devices.filter(i => i.kind === 'audioinput').map(toDevice)
     const videos = devices.filter(i => i.kind === 'videoinput').map(toDevice)
 
+    if ( currentDeviceSpeaker === deviceNone.deviceId) {
+      let device = speakers[0];
+      if (device) await setCurrentDeviceSpeaker(device.deviceId);
+    } 
+
     if (currentDeviceAudio === deviceNone.deviceId) {
       const device = audios[0]
       if (device) await setCurrentDeviceAudio(device.deviceId)
@@ -95,6 +107,7 @@ export default function DeviceBar(props: { streamId: string }) {
       if (device) await setCurrentDeviceVideo(device.deviceId)
     }
 
+    setDeviceSpeaker([...speakers])
     setDeviceAudio([...audios])
     setDeviceVideo([...videos, deviceScreen])
   }
@@ -130,6 +143,12 @@ export default function DeviceBar(props: { streamId: string }) {
     return () => { navigator.mediaDevices.removeEventListener('devicechange', updateDeviceList) }
   }, [])
 
+  const onChangedDeviceSpeaker = async (current: string) => {
+    setLoadingSpeaker(true)
+    await setCurrentDeviceSpeaker(current)
+    setLoadingSpeaker(false)
+  }
+
   const onChangedDeviceAudio = async (current: string) => {
     setLoadingAudio(true)
     await setCurrentDeviceAudio(current)
@@ -151,6 +170,24 @@ export default function DeviceBar(props: { streamId: string }) {
   return (
     <div className="flex flex-row flex-wrap justify-around p-xs">
       <center className="flex flex-row flex-wrap justify-around">
+      <section className="m-1 p-1 flex flex-row justify-center rounded-md border-1 border-indigo-500">
+          <button className="text-rose-400 rounded-md w-8 h-8" onClick={async () => {
+          }}>
+            <center>{ loadingSpeaker ? <Loading/> : <SvgSpeaker/> }</center>
+          </button>
+          <div className="flex flex-col justify-between w-1 pointer-events-none">
+          </div>
+          <select
+            className="w-3.5 h-8 rounded-sm rotate-180"
+            value={currentDeviceSpeaker}
+            onChange={e => onChangedDeviceSpeaker(e.target.value)}
+          >
+            {deviceSpeaker.map(device =>
+              <option key={device.deviceId} value={device.deviceId}>{device.label}</option>
+            )}
+          </select>
+        </section>
+
         <section className="m-1 p-1 flex flex-row justify-center rounded-md border-1 border-indigo-500">
           <button className="text-rose-400 rounded-md w-8 h-8" onClick={async () => {
             setLoadingAudio(true)
diff --git a/webapp/components/svg/speaker.tsx b/webapp/components/svg/speaker.tsx
new file mode 100644
index 0000000..30c4087
--- /dev/null
+++ b/webapp/components/svg/speaker.tsx
@@ -0,0 +1,8 @@
+export default function SvgSpeaker() {
+  return (
+    <svg focusable="false" width="24" height="24" viewBox="0 0 24 24">
+      <path fill="currentColor" d="M4 7h4l5-5v20l-5-5H4c-1.1 0-2-.9-2-2V9c0-1.1.9-2 2-2zm11.54 3.88l1.41-1.41c1.78 1.78 1.78 4.66 0 6.44l-1.41-1.41c1.17-1.17 1.17-3.07 0-4.24zm2.83-2.83l1.41-1.41c3.12 3.12 3.12 8.19 0 11.31l-1.41-1.41c2.34-2.34 2.34-6.13 0-8.49z"></path>
+    </svg>
+  );
+}
+  
\ No newline at end of file
diff --git a/webapp/components/use/whip.ts b/webapp/components/use/whip.ts
index d63b8db..e144a42 100644
--- a/webapp/components/use/whip.ts
+++ b/webapp/components/use/whip.ts
@@ -13,25 +13,32 @@ interface WHIPData extends Data {
   setUserName: (name: string) => void,
   setSyncUserStatus: (callback: (userStatus: Stream) => void) => void,
 
+  currentDeviceSpeaker: string,
   currentDeviceAudio: string,
   currentDeviceVideo: string,
+  setCurrentDeviceSpeaker: (current: string) => Promise<void>,
   setCurrentDeviceAudio: (current: string) => Promise<void>,
   setCurrentDeviceVideo: (current: string) => Promise<void>,
+  toggleEnableSpeaker: () => Promise<void>,
   toggleEnableAudio: () => Promise<void>,
   toggleEnableVideo: () => Promise<void>,
 }
 
 class WHIPContext extends Context {
+  private audioElement: HTMLAudioElement
   client: WHIPClient = new WHIPClient()
   cache: WHIPData
 
+  currentDeviceSpeaker = deviceNone.deviceId
   currentDeviceAudio = deviceNone.deviceId
   currentDeviceVideo = deviceNone.deviceId
+  toggleEnableSpeaker = async () => this.setCurrentDeviceSpeaker(this.userStatus.speaker ? deviceNone.deviceId : this.currentDeviceSpeaker)
   toggleEnableAudio = async () => this.setCurrentDeviceAudio(this.userStatus.audio ? deviceNone.deviceId : this.currentDeviceAudio)
   toggleEnableVideo = async () => this.setCurrentDeviceVideo(this.userStatus.video ? deviceNone.deviceId : this.currentDeviceVideo)
 
   constructor(id: string) {
     super(id)
+    this.audioElement = document.createElement("audio")
     this.cache = this.clone()
   }
 
@@ -64,10 +71,13 @@ class WHIPContext extends Context {
       setUserName: (name: string) => this.setUserName(name),
       setSyncUserStatus: (callback: (userStatus: Stream) => void) => this.setSyncUserStatus(callback),
 
+      currentDeviceSpeaker: this.currentDeviceSpeaker,
       currentDeviceAudio: this.currentDeviceAudio,
       currentDeviceVideo: this.currentDeviceVideo,
+      setCurrentDeviceSpeaker: (current: string) => this.setCurrentDeviceSpeaker(current),
       setCurrentDeviceAudio: (current: string) => this.setCurrentDeviceAudio(current),
       setCurrentDeviceVideo: (current: string) => this.setCurrentDeviceVideo(current),
+      toggleEnableSpeaker: () => this.toggleEnableSpeaker(),
       toggleEnableAudio: () => this.toggleEnableAudio(),
       toggleEnableVideo: () => this.toggleEnableVideo(),
     }
@@ -98,6 +108,39 @@ class WHIPContext extends Context {
     }
   }
 
+  async setCurrentDeviceSpeaker(current: string) {
+    const { userStatus, currentDeviceSpeaker } = this;
+
+      // 检查是否需要切换设备
+      if (current !== currentDeviceSpeaker || !userStatus.speaker) {
+        if (!this.audioElement) {
+          // 使用全局或共享的 audioElement
+          this.audioElement = document.createElement("audio");
+          document.body.appendChild(this.audioElement);
+        }
+  
+    // 设置新的扬声器设备
+      if (this.audioElement.setSinkId) {
+        try {
+          await this.audioElement.setSinkId(current);
+          console.log(`扬声器切换到设备: ${current}`);
+        } catch (error) {
+          console.error("切换音频输出设备失败:", error);
+        }
+      } else {
+        console.error("当前浏览器不支持 setSinkId");
+      }
+  
+        // 更新状态
+        userStatus.speaker = current === deviceNone.deviceId ? false : true
+        this.currentDeviceSpeaker = current === deviceNone.deviceId ? this.currentDeviceSpeaker : current
+  
+        // 同步状态并触发相关回调
+        this.sync();
+        this.syncUserStatus(userStatus);
+      }
+  }
+
   onChangedDeviceAudio() {
     const { pc, stream } = this
     // If WebRTC is connected, switch track
diff --git a/webapp/components/use/whxp.ts b/webapp/components/use/whxp.ts
index 8de1d9f..bcbc5e7 100644
--- a/webapp/components/use/whxp.ts
+++ b/webapp/components/use/whxp.ts
@@ -19,6 +19,7 @@ class Context extends EventTarget {
   userStatus: Stream = {
     name: '',
     state: StreamState.New,
+    speaker: true,
     audio: true,
     video: true,
     screen: false,
diff --git a/webapp/lib/api.ts b/webapp/lib/api.ts
index ca72dae..8cf47a6 100644
--- a/webapp/lib/api.ts
+++ b/webapp/lib/api.ts
@@ -24,6 +24,7 @@ enum StreamState {
 interface Stream {
   name: string,
   state: StreamState
+  speaker: boolean,
   audio: boolean,
   video: boolean,
   screen: boolean,