diff --git a/.vscode/settings.json b/.vscode/settings.json
index ec42c3f..c35668e 100644
--- a/.vscode/settings.json
+++ b/.vscode/settings.json
@@ -30,5 +30,14 @@
     "editor.defaultFormatter": "esbenp.prettier-vscode",
     "editor.maxTokenizationLineLength": 2500
   },
+  "[rust]": {
+    "editor.quickSuggestions": {
+      "other": true,
+      "comments": false,
+      "strings": true
+    },
+    "editor.tabSize": 4,
+    "editor.formatOnSave": true,
+  },
   "vue3snippets.enable-compile-vue-file-on-did-save-code": false
 }
\ No newline at end of file
diff --git a/README.md b/README.md
index b3ecaee..27cc5d5 100644
--- a/README.md
+++ b/README.md
@@ -1,3 +1,13 @@
 Contents edited from <https://www.howtostudykorean.com> is at [/docs/howtostudykorean/](/docs/howtostudykorean/).
 
 For specifics of how this is created, see <https://git.polv.cc/polv/zh-grammar#how-it-works-for-android>
+
+## TTS
+
+For Windows TTS, I don't like Korean language support pack sound, so I use
+
+```shell
+docker run -it -p 5500:5500 synesthesiam/opentts:ko
+```
+
+It's required to have Docker running.
diff --git a/docs/.vitepress/theme/components/TTS.vue b/docs/.vitepress/theme/components/TTS.vue
index f6dba37..6f829ea 100644
--- a/docs/.vitepress/theme/components/TTS.vue
+++ b/docs/.vitepress/theme/components/TTS.vue
@@ -3,6 +3,7 @@
     <span ref="main" :class="class" :lang="lang">
       <slot />
     </span>
+    <audio style="display: none" ref="audio"></audio>
     <span>
       <button type="button" @click="playAudio()">🗣️</button>
     </span>
@@ -23,12 +24,37 @@ const { lang = "ko-KR" } = defineProps<{
 }>();
 
 const mainEl = useTemplateRef("main");
+const audioEl = useTemplateRef("audio");
+
+function getMainText() {
+  if (!mainEl.value) return "";
+
+  return mainEl.value.innerText
+    .replace(/[*]/gu, " ")
+    .replace(/[\p{Ps}\p{Pe}]/gu, "");
+}
 
 async function playAudio() {
   if (!mainEl.value) return;
 
+  if (navigator.platform === "Win32" || !(await playAudioTTSPlugin())) {
+    if (!audioEl.value) return;
+
+    const url = new URL("http://localhost:5500/api/tts");
+    url.searchParams.set("voice", "glow-speak:ko_kss");
+    url.searchParams.set("text", getMainText());
+
+    audioEl.value.src = url.href;
+    audioEl.value.currentTime = 0;
+    audioEl.value.play();
+  }
+}
+
+async function playAudioTTSPlugin(): Promise<boolean> {
+  if (!mainEl.value) return false;
+
   const voices = await getVoices(lang);
-  if (!voices.length) return;
+  if (!voices.length) return false;
 
   document.querySelectorAll("audio").forEach((el) => {
     el.pause();
@@ -37,9 +63,7 @@ async function playAudio() {
   const v = voices[Math.floor(Math.random() * voices.length)];
   console.log(JSON.stringify(v));
   await speak({
-    text: mainEl.value.innerText
-      .replace(/[*]/gu, " ")
-      .replace(/[\p{Ps}\p{Pe}]/gu, ""),
+    text: getMainText(),
     language: null,
     rate: null,
     volume: null,
@@ -47,6 +71,8 @@ async function playAudio() {
     queueMode: null,
     voiceId: v.id,
   });
+
+  return true;
 }
 </script>