Update app.py

app.py

@@ -35,6 +35,7 @@ DEFAULT_SYSTEM_PROMPT = (
     "Generate audio following instruction.\n\n"
     "<|scene_desc_start|>\n"
     "Audio is recorded from a quiet room.\n"
+    "Support for multiple languages including English, Chinese, Korean, Japanese, and more.\n"
     "<|scene_desc_end|>"
 )
 
@@ -99,6 +100,19 @@ PREDEFINED_EXAMPLES = {
         "icon": "🇨🇳",
         "color": "#AA96DA"
     },
+    "single-speaker-kr": {
+        "system_prompt": "Generate audio following instruction.\n\n"
+        "<|scene_desc_start|>\n"
+        "Audio is recorded from a quiet room.\n"
+        "<|scene_desc_end|>",
+        "input_text": "안녕하세요, 오늘은 인공지능의 미래에 대해 이야기해보겠습니다.\n"
+        "최근 AI 기술의 발전이 정말 놀라운데요,\n"
+        "특히 음성 합성 기술은 이제 사람과 구별하기 어려울 정도로 자연스러워졌습니다.\n"
+        "여러분은 제가 실제 사람인지 AI인지 구별할 수 있으신가요?",
+        "description": "🇰🇷 <b>Korean Speech</b> - Generate natural Korean speech",
+        "icon": "🇰🇷",
+        "color": "#FFB6C1"
+    },
     "single-speaker-bgm": {
         "system_prompt": DEFAULT_SYSTEM_PROMPT,
         "input_text": "[music start] I will remember this, thought Ender, when I am defeated. To keep dignity, and give honor where it's due, so that defeat is not disgrace. And I hope I don't have to do it often. [music end]",
@@ -196,8 +210,18 @@ def normalize_chinese_punctuation(text):
 
 
 def normalize_text(transcript: str):
+    # Skip normalization for Korean text to preserve it properly
+    if any('\u3131' <= char <= '\u3163' or '\uac00' <= char <= '\ud7a3' for char in transcript):
+        # Korean text detected - minimal normalization
+        transcript = transcript.strip()
+        if transcript and not any([transcript.endswith(c) for c in [".", "!", "?", "。", "！", "？"]]):
+            transcript += "."
+        return transcript
+    
+    # Chinese punctuation normalization
     transcript = normalize_chinese_punctuation(transcript)
-    # Other normalizations (e.g., parentheses and other symbols. Will be improved in the future)
+    
+    # Other normalizations (e.g., parentheses and other symbols)
     transcript = transcript.replace("(", " ")
     transcript = transcript.replace(")", " ")
     transcript = transcript.replace("°F", " degrees Fahrenheit")