From 9bca8880fef1607fe16e20ad06124e292dc739d0 Mon Sep 17 00:00:00 2001
From: Yueqian Lin <70319226+linyueqian@users.noreply.github.com>
Date: Tue, 27 Jan 2026 10:19:51 -0500
Subject: [PATCH] Fix TTS speaker typo and add supported languages (#990)

Signed-off-by: linyueqian <linyueqian@outlook.com>
---
 pyproject.toml                                   |  1 +
 .../openai_api/test_serving_speech.py            |  2 +-
 vllm_omni/entrypoints/openai/serving_speech.py   | 16 ++++++++++++++--
 3 files changed, 16 insertions(+), 3 deletions(-)

diff --git a/pyproject.toml b/pyproject.toml
index 279d8e7d3..306733063 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -200,4 +200,5 @@ extend-ignore-identifiers-re = [
     ".*nothink.*",
     ".*NOTHINK.*",
     ".*nin.*",
+    "Ono_Anna",
 ]
diff --git a/tests/entrypoints/openai_api/test_serving_speech.py b/tests/entrypoints/openai_api/test_serving_speech.py
index 82b5a4fef..675af4141 100644
--- a/tests/entrypoints/openai_api/test_serving_speech.py
+++ b/tests/entrypoints/openai_api/test_serving_speech.py
@@ -308,7 +308,7 @@ class TestTTSMethods:
         assert speech_server._validate_tts_request(req) == "Input text cannot be empty"
 
         # Invalid language
-        req = OpenAICreateSpeechRequest(input="Hello", language="French")
+        req = OpenAICreateSpeechRequest(input="Hello", language="InvalidLang")
         assert "Invalid language" in speech_server._validate_tts_request(req)
 
         # Invalid speaker
diff --git a/vllm_omni/entrypoints/openai/serving_speech.py b/vllm_omni/entrypoints/openai/serving_speech.py
index d5bc20cde..223557aab 100644
--- a/vllm_omni/entrypoints/openai/serving_speech.py
+++ b/vllm_omni/entrypoints/openai/serving_speech.py
@@ -27,10 +27,22 @@ _TTS_SPEAKERS: set[str] = {
     "Eric",
     "Ryan",
     "Aiden",
-    "One_Anna",
+    "Ono_Anna",
     "Sohee",
 }
-_TTS_LANGUAGES: set[str] = {"Auto", "Chinese", "English", "Japanese", "Korean"}
+_TTS_LANGUAGES: set[str] = {
+    "Auto",
+    "Chinese",
+    "English",
+    "Japanese",
+    "Korean",
+    "German",
+    "French",
+    "Russian",
+    "Portuguese",
+    "Spanish",
+    "Italian",
+}
 _TTS_MAX_INSTRUCTIONS_LENGTH = 500
 _TTS_MAX_NEW_TOKENS_MIN = 1
 _TTS_MAX_NEW_TOKENS_MAX = 4096