Skip to content

利用 AI 大模型和自动化工作流,根据主题或关键词一键生成高清短视频。Generate HD short videos from a topic or keyword with an automated AI workflow.

fix(tts): harden Kokoro voice and audio handling

harry0703883ef2f8d43f
diff --git a/app/services/voice.py b/app/services/voice.py
index 5d2efc6d..22c31401 100644
--- a/app/services/voice.py
+++ b/app/services/voice.py
@@ -224,19 +224,32 @@ def get_chatterbox_voices() -> list[str]:
KOKORO_DEFAULT_VOICE = "af_heart"
-def get_kokoro_voices() -> list[str]:
- """Return the Kokoro voices as ``kokoro:<name>`` entries.
+def _normalize_kokoro_voices(entries) -> list[str]:
+ """统一手工配置与服务端新旧格式,只接收真实 ID,避免把对象转成音色名。"""
+ if isinstance(entries, str):
+ entries = entries.split(",")
+ if not isinstance(entries, list):
+ return []
+ result = []
+ for entry in entries:
+ name = entry.get("id") if isinstance(entry, dict) else entry
+ if not isinstance(name, str):
+ continue
+ name = name.strip().removeprefix("kokoro:").strip()
+ if name:
+ value = f"kokoro:{name}"
+ if value not in result:
+ result.append(value)
+ return result
- Operators may pin the list via ``[kokoro] voices`` (a TOML array, or a
- comma-separated string). When it is empty the server is asked:
- Kokoro-FastAPI and compatible servers expose ``GET {base_url}/audio/voices``
- → ``{"voices": [...]}``. If the server cannot be reached the dropdown falls
- back to Kokoro's default voice so the UI stays usable.
+
+def get_kokoro_voices(*, fallback: bool = True) -> list[str]:
+ """手工音色优先,否则查询服务器;UI 可禁用默认值以识别断线并保留选择。
+
+ 旧版服务器返回字符串列表,新版返回包含 id 的对象列表。失败时不在服务层
+ 缓存会话状态;由 WebUI 保留上次成功目录,避免不同用户/端点互相污染。
"""
- voices = config.kokoro.get("voices", []) or []
- if isinstance(voices, str):
- voices = [v.strip() for v in voices.split(",") if v.strip()]
- voices = [str(v).strip() for v in voices if str(v).strip()]
+ voices = _normalize_kokoro_voices(config.kokoro.get("voices"))
if not voices:
base_url = (config.kokoro.get("base_url", "") or "").strip().rstrip("/")
if base_url:
@@ -251,17 +264,17 @@ def get_kokoro_voices() -> list[str]:
if response.status_code == 200:
data = response.json()
listed = data.get("voices", []) if isinstance(data, dict) else data
- voices = [str(v).strip() for v in (listed or []) if str(v).strip()]
+ voices = _normalize_kokoro_voices(listed)
+ if not voices:
+ logger.warning("kokoro voice list contains no valid voice IDs")
else:
logger.warning(
f"kokoro voices request failed with status {response.status_code}"
)
except Exception as e:
- logger.warning(f"kokoro voices unavailable at {base_url}: {str(e)}")
- result = [v if v.startswith("kokoro:") else f"kokoro:{v}" for v in voices]
- if not result:
- result = [f"kokoro:{KOKORO_DEFAULT_VOICE}"]
- return result
+ # 不输出 URL/异常正文,避免自托管地址的查询参数或认证信息进入日志。
+ logger.warning(f"kokoro voice list unavailable ({type(e).__name__})")
+ return voices or ([f"kokoro:{KOKORO_DEFAULT_VOICE}"] if fallback else [])
def get_fish_audio_voices() -> list[str]:
@@ -1744,12 +1757,11 @@ def _openai_compatible_tts(
# 1.0-centred multiplier, so it maps directly (clamped to the valid range).
"speed": max(0.25, min(4.0, float(voice_rate or 1.0))),
}
- # voice_volume is accepted by the callers for parity with the other TTS
- # providers but is intentionally not sent: the OpenAI /audio/speech contract
- # has no volume field, so these servers ignore it. Adjust loudness via
- # voice_rate (speed) or in post-processing instead.
+ # OpenAI speech 协议没有音量字段;最终混音阶段应用 voice_volume。
+ # speed 仅调节语速,不能作为音量参数使用。
for i in range(3):
+ temporary_audio = None
try:
logger.info(f"start {provider} tts, voice: {voice}, try: {i + 1}")
ensure_file_path_exists(voice_file)
@@ -1761,16 +1773,28 @@ def _openai_compatible_tts(
)
continue
- with open(voice_file, "wb") as f:
+ if not response.content:
+ raise ValueError(f"{provider} returned empty audio")
+
+ # 先写入同目录临时文件并真实解码,成功后才替换目标。失败不能破坏
+ # 已有试听/配音;先关闭文件再解码和替换,兼容 Windows 文件占用规则。
+ with tempfile.NamedTemporaryFile(
+ dir=os.path.dirname(os.path.abspath(voice_file)),
+ suffix=".mp3", delete=False,
+ ) as f:
+ temporary_audio = f.name
f.write(response.content)
- audio_clip = AudioFileClip(voice_file)
+ audio_clip = AudioFileClip(temporary_audio)
try:
audio_duration = audio_clip.duration
finally:
audio_clip.close()
+ if not math.isfinite(audio_duration) or audio_duration <= 0:
+ raise ValueError(f"{provider} returned an invalid audio duration")
sub_maker = ensure_legacy_submaker_fields(SubMaker())
+ os.replace(temporary_audio, voice_file)
logger.success(f"{provider} tts succeeded: {voice_file}")
return populate_legacy_submaker_with_full_text(
sub_maker=sub_maker,
@@ -1779,6 +1803,13 @@ def _openai_compatible_tts(
)
except Exception as e:
logger.error(f"{provider} tts failed: {str(e)}")
+ finally:
+ if temporary_audio and os.path.exists(temporary_audio):
+ try:
+ os.unlink(temporary_audio)
+ except OSError as exc:
+ # 清理失败仍保持 TTS 的失败返回约定,不用次要异常掩盖原始原因。
+ logger.warning(f"could not remove temporary {provider} audio: {exc}")
return None
@@ -1855,6 +1886,11 @@ def kokoro_tts(
if not text:
logger.error("Kokoro TTS text is empty")
return None
+ # 纯标点/表情没有可发音文字;真实服务可能以 HTTP 200 返回空 MP3,
+ # 提前终止可避免无效请求及 MoviePy 在解码空文件时的底层异常。
+ if not any(character.isalnum() for character in text):
+ logger.error("Kokoro TTS text contains no speakable characters")
+ return None
base_url = (config.kokoro.get("base_url", "") or "").strip().rstrip("/")
if not base_url:
logger.error(
diff --git a/test/services/test_kokoro.py b/test/services/test_kokoro.py
new file mode 100644
index 00000000..f3a086a4
--- /dev/null
+++ b/test/services/test_kokoro.py
@@ -0,0 +1,133 @@
+"""Kokoro 协议兼容和共享音频传输回归,不依赖外部服务。"""
+
+from types import SimpleNamespace
+from unittest.mock import Mock
+
+import pytest
+import requests
+
+from app.config import config
+from app.services import voice
+
+
+def kokoro_config(monkeypatch):
+ settings = {"base_url": "http://localhost:8880/v1/", "api_key": "test-key",
+ "model_id": "kokoro", "voices": []}
+ monkeypatch.setattr(config, "kokoro", settings)
+ return settings
+
+
[email protected]("payload", [
+ {"voices": ["af_heart", "zf_xiaobei"]},
+ {"voices": [{"id": "af_heart", "name": "Heart"}, {"id": "zf_xiaobei"}]},
+ ["af_heart", {"id": "zf_xiaobei"}],
+])
+def test_voice_list_supports_old_and_new_servers(monkeypatch, kokoro_config, payload):
+ get = Mock(return_value=SimpleNamespace(status_code=200, json=lambda: payload))
+ monkeypatch.setattr(voice.requests, "get", get)
+ assert voice.get_kokoro_voices() == ["kokoro:af_heart", "kokoro:zf_xiaobei"]
+ get.assert_called_once_with("http://localhost:8880/v1/audio/voices",
+ headers={"Authorization": "Bearer test-key"}, timeout=5)
+
+
[email protected]("entries, expected", [
+ ([" af_heart ", "kokoro:af_heart", {"id": " zf_xiaobei "}, None, 2,
+ {"name": "invalid"}, {"id": None}, "kokoro:"], ["kokoro:af_heart", "kokoro:zf_xiaobei"]),
+ (" af_heart, kokoro:zf_xiaobei, ", ["kokoro:af_heart", "kokoro:zf_xiaobei"]),
+ ({"id": "af_heart"}, []), (42, []), (None, []),
+])
+def test_voice_normalization_rejects_invalid_entries(entries, expected):
+ assert voice._normalize_kokoro_voices(entries) == expected
+
+
+def test_pinned_voices_do_not_request_server(monkeypatch, kokoro_config):
+ kokoro_config["voices"] = "af_heart, zf_xiaobei"
+ get = Mock(side_effect=AssertionError("manual voices must not request server"))
+ monkeypatch.setattr(voice.requests, "get", get)
+ assert voice.get_kokoro_voices(fallback=False) == ["kokoro:af_heart", "kokoro:zf_xiaobei"]
+ get.assert_not_called()
+
+
[email protected]("failure", [requests.Timeout(), requests.ConnectionError(),
+ ValueError("invalid JSON"), 401, 500, [], {"voices": None}])
+def test_voice_discovery_failure_is_distinguishable(monkeypatch, kokoro_config, failure):
+ if isinstance(failure, Exception):
+ get = Mock(side_effect=failure)
+ else:
+ get = Mock(return_value=SimpleNamespace(
+ status_code=failure if isinstance(failure, int) else 200,
+ json=lambda: failure,
+ ))
+ monkeypatch.setattr(voice.requests, "get", get)
+ assert voice.get_kokoro_voices(fallback=False) == []
+ assert voice.get_kokoro_voices() == ["kokoro:af_heart"]
+
+
[email protected]("text", ["", " ", "...!!!", "😀"])
+def test_unspeakable_text_does_not_make_requests(monkeypatch, kokoro_config, tmp_path, text):
+ post = Mock()
+ monkeypatch.setattr(voice.requests, "post", post)
+ assert voice.kokoro_tts(text, "af_heart", str(tmp_path / "test.mp3")) is None
+ post.assert_not_called()
+
+
[email protected]("provider", ["kokoro", "chatterbox"])
[email protected]("rate, expected", [(0.1, 0.25), (1.2, 1.2), (5, 4.0)])
+def test_transport_closes_audio_and_preserves_contract(monkeypatch, tmp_path, provider, rate, expected):
+ monkeypatch.setattr(config, provider, {
+ "base_url": "http://localhost:8880/v1/", "api_key": "key", "model_id": provider,
+ })
+ post = Mock(return_value=SimpleNamespace(status_code=200, content=b"audio", text=""))
+ monkeypatch.setattr(voice.requests, "post", post)
+ clip = Mock(duration=1.5)
+ monkeypatch.setattr(voice, "AudioFileClip", Mock(return_value=clip))
+ output = tmp_path / "output.mp3"
+ maker = voice.tts("Hello world.", f"{provider}:af_heart-Female", rate, str(output))
+ assert maker is not None
+ assert output.read_bytes() == b"audio"
+ assert list(tmp_path.iterdir()) == [output]
+ clip.close.assert_called_once()
+ payload = post.call_args.kwargs["json"]
+ assert payload == {"model": provider, "input": "Hello world.", "voice": "af_heart",
+ "response_format": "mp3", "speed": expected}
+ assert post.call_args.kwargs["headers"]["Authorization"] == "Bearer key"
+
+
[email protected]("provider", ["kokoro", "chatterbox"])
[email protected]("failure", ["empty", "decode", "zero", "nan", "replace", "http", "timeout"])
+def test_failed_audio_never_overwrites_output(monkeypatch, tmp_path, provider, failure):
+ """失败保留原文件,解码资源与临时文件均释放,包括 Windows 替换失败。"""
+ output = tmp_path / "existing.mp3"
+ output.write_bytes(b"previous audio")
+ post = Mock(return_value=SimpleNamespace(
+ status_code=503 if failure == "http" else 200,
+ content=b"" if failure == "empty" else b"invalid audio", text="Unavailable"))
+ if failure == "timeout":
+ post.side_effect = requests.Timeout()
+ monkeypatch.setattr(voice.requests, "post", post)
+ clip = Mock(duration=0 if failure == "zero" else float("nan") if failure == "nan" else 1)
+ reader = Mock(return_value=clip)
+ if failure == "decode":
+ reader.side_effect = ValueError("invalid audio")
+ monkeypatch.setattr(voice, "AudioFileClip", reader)
+ if failure == "replace":
+ monkeypatch.setattr(voice.os, "replace", Mock(side_effect=PermissionError("in use")))
+ assert voice._openai_compatible_tts(
+ provider, "http://localhost/v1", "", provider, "af_heart", "Hello", 1, str(output)
+ ) is None
+ assert output.read_bytes() == b"previous audio"
+ assert list(tmp_path.iterdir()) == [output]
+ assert post.call_count == 3
+ if failure in {"zero", "nan", "replace"}:
+ assert clip.close.call_count == 3
+
+
+def test_transient_error_retries_successfully(monkeypatch, tmp_path):
+ responses = [requests.Timeout(), SimpleNamespace(status_code=200, content=b"audio", text="")]
+ post = Mock(side_effect=responses)
+ monkeypatch.setattr(voice.requests, "post", post)
+ monkeypatch.setattr(voice, "AudioFileClip", Mock(return_value=Mock(duration=1)))
+ assert voice._openai_compatible_tts("kokoro", "http://localhost/v1", "", "kokoro",
+ "af_heart", "Hello", 1, str(tmp_path / "a.mp3"))
+ assert post.call_count == 2
diff --git a/test/services/test_kokoro_integration.py b/test/services/test_kokoro_integration.py
new file mode 100644
index 00000000..769aba8c
--- /dev/null
+++ b/test/services/test_kokoro_integration.py
@@ -0,0 +1,50 @@
+"""显式启用的真实服务测试;普通 CI 不需要安装或访问 Kokoro。
+
+启动本机服务后,设置 MPT_KOKORO_TEST_BASE_URL=http://127.0.0.1:8880/v1
+运行本文件。测试只合成短文本,不调用 LLM、素材平台或发布接口。
+"""
+
+import os
+
+import pytest
+
+from app.config import config
+from app.services import voice
+
+
+BASE_URL = os.environ.get("MPT_KOKORO_TEST_BASE_URL", "")
+pytestmark = pytest.mark.skipif(not BASE_URL, reason="requires an explicitly configured Kokoro test server")
+
+
+def live_config(monkeypatch):
+ monkeypatch.setattr(config, "kokoro", {"base_url": BASE_URL, "api_key": "",
+ "model_id": "kokoro", "voices": []})
+
+
+def test_live_voice_discovery(live_config):
+ options = voice.get_kokoro_voices(fallback=False)
+ assert "kokoro:af_heart" in options
+ assert "kokoro:zf_xiaobei" in options
+ assert all(option.startswith("kokoro:") and "{" not in option for option in options)
+
+
[email protected]("speaker, text", [
+ ("af_heart", "Artificial intelligence is changing everyday life."),
+ ("zf_xiaobei", "人工智能正在改变我们的生活。合理使用技术,让每个人受益。"),
+ ("ff_siwis", "Bonjour ! L'intelligence artificielle nous aide à apprendre."),
+ ("ef_dora", "La inteligencia artificial mejora nuestra vida cotidiana."),
+ ("if_sara", "L'intelligenza artificiale migliora la vita di ogni giorno."),
+ ("pf_dora", "A inteligência artificial melhora a nossa vida."),
+ ("hf_alpha", "नमस्ते। कृत्रिम बुद्धिमत्ता हमारे जीवन को बेहतर बनाती है।"),
+])
+def test_live_audio_and_subtitles(live_config, tmp_path, speaker, text):
+ output = tmp_path / "audio.mp3"
+ maker = voice.tts(text, f"kokoro:{speaker}", 1, str(output))
+ assert maker is not None
+ assert output.stat().st_size > 100
+ assert voice.get_audio_duration(str(output)) > 0
+ subtitle = tmp_path / "subtitle.srt"
+ voice.create_subtitle(maker, text, str(subtitle))
+ assert "-->" in subtitle.read_text(encoding="utf-8")
+ assert sorted(path.suffix for path in tmp_path.iterdir()) == [".mp3", ".srt"]
diff --git a/test/services/test_webui_kokoro.py b/test/services/test_webui_kokoro.py
new file mode 100644
index 00000000..fd77acc3
--- /dev/null
+++ b/test/services/test_webui_kokoro.py
@@ -0,0 +1,88 @@
+"""运行实际 Streamlit 页面,验证 Kokoro 音色刷新、断线和配置切换。"""
+
+from pathlib import Path
+from types import SimpleNamespace
+from unittest.mock import Mock
+
+import pytest
+from streamlit.testing.v1 import AppTest
+
+from app.config import config
+from app.services import voice
+
+
+def ui(monkeypatch):
+ monkeypatch.setattr(config, "ui", dict(config.ui, voice_mode="tts", tts_server="kokoro",
+ voice_name="kokoro:zf_xiaobei"))
+ monkeypatch.setattr(config, "kokoro", dict(base_url="http://localhost:8880/v1",
+ api_key="", model_id="kokoro", voices=[]))
+ monkeypatch.setattr(config, "save_config", Mock())
+ monkeypatch.setattr(config, "try_save_config", Mock(return_value=True))
+ page = AppTest.from_file(str(Path(__file__).parents[2] / "webui/Main.py"), default_timeout=30)
+ page.session_state["ui_language"] = "zh"
+ return page
+
+
+def selected_voice(page):
+ return next(item for item in page.selectbox
+ if str(item.key).startswith("speech_synthesis_select_kokoro"))
+
+
+def test_online_offline_recovery_retains_selection(monkeypatch, ui):
+ """真实 rerun 穿过一次成功、一次断线、恢复,不能把保存音色改为默认值。"""
+ get = Mock(side_effect=[["kokoro:af_heart", "kokoro:zf_xiaobei"], [],
+ ["kokoro:af_heart", "kokoro:zf_xiaobei"]])
+ monkeypatch.setattr(voice, "get_kokoro_voices", get)
+ ui.run()
+ assert selected_voice(ui).value == "kokoro:zf_xiaobei"
+ ui.run()
+ assert get.call_count == 1 # 缓存期内不重复访问服务。
+ ui.session_state["kokoro_voice_catalog"]["checked_at"] = -100
+ ui.run()
+ assert selected_voice(ui).value == "kokoro:zf_xiaobei"
+ assert config.ui["voice_name"] == "kokoro:zf_xiaobei"
+ assert any("暂时无法获取 Kokoro" in str(w.value) for w in ui.warning)
+ ui.session_state["kokoro_voice_catalog"]["checked_at"] = -100
+ ui.run()
+ assert selected_voice(ui).value == "kokoro:zf_xiaobei"
+ assert not any("暂时无法获取 Kokoro" in str(w.value) for w in ui.warning)
+ assert get.call_count == 3
+ assert not ui.exception
+
+
+def test_first_open_offline_preserves_saved_voice(monkeypatch, ui):
+ monkeypatch.setattr(voice, "get_kokoro_voices", Mock(return_value=[]))
+ ui.run()
+ assert selected_voice(ui).value == "kokoro:zf_xiaobei"
+ assert config.ui["voice_name"] == "kokoro:zf_xiaobei"
+ assert not ui.exception
+
+
[email protected]("input_key, value", [
+ ("kokoro_base_url_input", "http://localhost:9999/v1"),
+ ("kokoro_api_key_input", "new-test-key"),
+])
+def test_endpoint_or_credential_change_invalidates_cache(monkeypatch, ui, input_key, value):
+ get = Mock(return_value=["kokoro:zf_xiaobei"])
+ monkeypatch.setattr(voice, "get_kokoro_voices", get)
+ ui.run()
+ next(item for item in ui.text_input if item.key == input_key).set_value(value).run()
+ assert get.call_count == 2
+ assert not ui.exception
+
+
+def test_manual_voices_and_clearing_them(monkeypatch, ui):
+ """手填音色立即使用;清空后自动发现,新版对象列表显示为简洁 ID。"""
+ config.kokoro["voices"] = ["zf_xiaobei", "af_heart"]
+ get = Mock(return_value=SimpleNamespace(status_code=200, json=lambda: {
+ "voices": [{"id": "zf_xiaobei"}, {"id": "af_heart"}],
+ }))
+ monkeypatch.setattr(voice.requests, "get", get)
+ ui.run()
+ get.assert_not_called()
+ next(item for item in ui.text_input if item.key == "kokoro_voices_input").set_value("").run()
+ get.assert_called_once()
+ assert selected_voice(ui).options == ["zf_xiaobei", "af_heart"]
+ assert selected_voice(ui).value == "kokoro:zf_xiaobei"
+ assert not ui.exception
diff --git a/webui/Main.py b/webui/Main.py
index 93704f28..684d98e2 100644
--- a/webui/Main.py
+++ b/webui/Main.py
@@ -434,11 +434,8 @@ def _sync_chatterbox_config_from_session_state():
def _sync_kokoro_config_from_session_state():
- # 与 Chatterbox 相同:Kokoro 配置输入框位于试听按钮之后,先从 session_state 同步一次。
- # Streamlit 的按钮会触发整页 rerun,而 Kokoro 配置输入框位于
- # “试听语音合成”按钮之后。如果试听时只读取 config.kokoro,可能拿不到
- # 用户刚在输入框里填入的 base_url/model/voices。先从 session_state 同步一次,
- # 可以保证按钮逻辑和输入框显示逻辑使用同一份最新配置。
+ # 音色目录先于设置输入框渲染,先同步浏览器状态,确保本次 rerun 就使用
+ # 新端点和手工音色配置,不必再操作一次控件。
_set_runtime_config(
"kokoro",
"base_url",
@@ -480,6 +477,38 @@ def _sync_kokoro_config_from_session_state():
)
+def _get_kokoro_voice_options(saved_voice_name: str) -> list[str]:
+ """会话内短缓存远端目录,断线时保留上次选择,不把故障当成用户改音色。"""
+ if config.kokoro.get("voices"):
+ return voice.get_kokoro_voices()
+
+ # 仅保留当前服务的一条缓存。更换端点/凭据立即重查,缓存不保存明文 Key;
+ # 30 秒内的其他 UI 操作不重复阻塞 5 秒等待一个已知离线的服务。
+ signature = (
+ (config.kokoro.get("base_url") or "").strip().rstrip("/"),
+ _credential_signature(config.kokoro.get("api_key", "")),
+ )
+ catalog = st.session_state.get("kokoro_voice_catalog", {})
+ if catalog.get("signature") != signature:
+ catalog = {"signature": signature, "voices": [], "checked_at": None}
+ now = time.monotonic()
+ if catalog["checked_at"] is None or now - catalog["checked_at"] >= 30:
+ fetched = voice.get_kokoro_voices(fallback=False)
+ catalog.update(checked_at=now, available=bool(fetched))
+ if fetched:
+ catalog["voices"] = fetched
+ st.session_state["kokoro_voice_catalog"] = catalog
+
+ options = list(catalog["voices"])
+ if not catalog["available"]:
+ st.warning(tr("Kokoro Voices Unavailable"))
+ # 首次打开时可能没有缓存,仍保留配置文件中的真实选择;恢复连接后
+ # 只有成功返回的新目录才能判定某个旧音色确实已被服务器删除。
+ if voice.is_kokoro_voice(saved_voice_name) and saved_voice_name not in options:
+ options.insert(0, saved_voice_name)
+ return options or [f"kokoro:{voice.KOKORO_DEFAULT_VOICE}"]
+
+
def _detect_audio_mime(audio_file: str, audio_bytes: bytes) -> str:
# 有些 OpenAI-compatible TTS 服务,例如 travisvn/chatterbox-tts-api,
# 即使请求 response_format=mp3,也会返回 WAV 内容。WebUI 试听如果固定
@@ -5950,7 +5979,7 @@ def _render_audio_settings(panel, params):
elif selected_tts_server == "kokoro":
# 自托管 Kokoro 服务的音色:[kokoro] voices 为空时从服务端 /audio/voices 读取
_sync_kokoro_config_from_session_state()
- filtered_voices = voice.get_kokoro_voices()
+ filtered_voices = _get_kokoro_voice_options(saved_voice_name)
elif selected_tts_server == "fish_audio":
filtered_voices = voice.get_fish_audio_voices()
else:
@@ -5974,7 +6003,7 @@ def _render_audio_settings(panel, params):
if voice.is_elevenlabs_voice(v):
parts = v.split(":", 2)
return parts[2] if len(parts) >= 3 else v
- if voice.is_chatterbox_voice(v):
+ if voice.is_chatterbox_voice(v) or voice.is_kokoro_voice(v):
name = v.split(":", 1)[1] if ":" in v else v
return name.replace("-Female", "").replace("-Male", "")
if voice.is_minimax_voice(v):
diff --git a/webui/i18n/de.json b/webui/i18n/de.json
index f0b81ed2..83b2c821 100644
--- a/webui/i18n/de.json
+++ b/webui/i18n/de.json
@@ -1,6 +1,13 @@
{
"Language": "Deutsch",
"Translation": {
+ "Kokoro Base URL": "Kokoro-Basis-URL",
+ "Kokoro API Key": "Kokoro-API-Schlüssel (optional; [Einrichtung](https://github.com/remsky/Kokoro-FastAPI))",
+ "Kokoro Model": "Kokoro-Modell",
+ "Kokoro Voices": "Kokoro-Stimmen",
+ "Kokoro Voices Placeholder": "Leer lassen, um Serverstimmen zu laden, z. B. af_heart, bf_emma, hf_alpha",
+ "Kokoro Voices Unavailable": "Die Kokoro-Stimmen konnten nicht geladen werden. Die bisherige Auswahl bleibt erhalten. Prüfe die Verbindung oder trage Stimmen manuell ein.",
+ "Kokoro Base URL Placeholder": "http://localhost:8880/v1",
"Login Required": "Anmeldung erforderlich",
"Please login to access settings": "Bitte melden Sie sich an, um auf die Einstellungen zuzugreifen",
"Username": "Benutzername",
diff --git a/webui/i18n/en.json b/webui/i18n/en.json
index 321bf23b..11d25d29 100644
--- a/webui/i18n/en.json
+++ b/webui/i18n/en.json
@@ -1,6 +1,7 @@
{
"Language": "English",
"Translation": {
+ "Kokoro Voices Unavailable": "Could not load Kokoro voices. Your previous selection is retained. Check the connection or enter voices manually.",
"Login Required": "Login Required",
"Please login to access settings": "Please login to access settings",
"Username": "Username",
diff --git a/webui/i18n/es.json b/webui/i18n/es.json
index 0df98f0c..eb3296f2 100644
--- a/webui/i18n/es.json
+++ b/webui/i18n/es.json
@@ -1,6 +1,13 @@
{
"Language": "Español",
"Translation": {
+ "Kokoro Base URL": "URL base de Kokoro",
+ "Kokoro API Key": "Clave API de Kokoro (opcional; [Configuración](https://github.com/remsky/Kokoro-FastAPI))",
+ "Kokoro Model": "Modelo de Kokoro",
+ "Kokoro Voices": "Voces de Kokoro",
+ "Kokoro Voices Placeholder": "Déjalo vacío para cargar las voces del servidor, o escribe af_heart, bf_emma, hf_alpha",
+ "Kokoro Voices Unavailable": "No se pudieron cargar las voces de Kokoro. Se conserva la selección anterior. Comprueba la conexión o introduce las voces manualmente.",
+ "Kokoro Base URL Placeholder": "http://localhost:8880/v1",
"Login Required": "Inicio de sesión requerido",
"Please login to access settings": "Inicia sesión para acceder a la configuración",
"Username": "Usuario",
diff --git a/webui/i18n/fr.json b/webui/i18n/fr.json
index e15cd13b..365217ea 100644
--- a/webui/i18n/fr.json
+++ b/webui/i18n/fr.json
@@ -1,6 +1,13 @@
{
"Language": "Français",
"Translation": {
+ "Kokoro Base URL": "URL de base de Kokoro",
+ "Kokoro API Key": "Clé API Kokoro (facultative ; [Configuration](https://github.com/remsky/Kokoro-FastAPI))",
+ "Kokoro Model": "Modèle Kokoro",
+ "Kokoro Voices": "Voix Kokoro",
+ "Kokoro Voices Placeholder": "Laisser vide pour charger les voix du serveur, ou saisir af_heart, bf_emma, hf_alpha",
+ "Kokoro Voices Unavailable": "Impossible de charger les voix Kokoro. La sélection précédente est conservée. Vérifiez la connexion ou saisissez les voix manuellement.",
+ "Kokoro Base URL Placeholder": "http://localhost:8880/v1",
"Login Required": "Connexion requise",
"Please login to access settings": "Veuillez vous connecter pour accéder aux paramètres",
"Username": "Nom d'utilisateur",
diff --git a/webui/i18n/id.json b/webui/i18n/id.json
index 1f0b788a..15d330ac 100644
--- a/webui/i18n/id.json
+++ b/webui/i18n/id.json
@@ -1,6 +1,13 @@
{
"Language": "Bahasa Indonesia",
"Translation": {
+ "Kokoro Base URL": "URL Dasar Kokoro",
+ "Kokoro API Key": "Kunci API Kokoro (opsional; [Panduan Konfigurasi](https://github.com/remsky/Kokoro-FastAPI))",
+ "Kokoro Model": "Model Kokoro",
+ "Kokoro Voices": "Suara Kokoro",
+ "Kokoro Voices Placeholder": "Kosongkan untuk memuat suara server, atau isi af_heart, bf_emma, hf_alpha",
+ "Kokoro Voices Unavailable": "Daftar suara Kokoro tidak dapat dimuat. Pilihan sebelumnya tetap disimpan. Periksa koneksi atau masukkan suara secara manual.",
+ "Kokoro Base URL Placeholder": "http://localhost:8880/v1",
"Login Required": "Login Diperlukan",
"Please login to access settings": "Silakan login untuk mengakses pengaturan",
"Username": "Nama Pengguna",
diff --git a/webui/i18n/it.json b/webui/i18n/it.json
index b653b5bd..8679d9f8 100644
--- a/webui/i18n/it.json
+++ b/webui/i18n/it.json
@@ -1,6 +1,13 @@
{
"Language": "Italiano",
"Translation": {
+ "Kokoro Base URL": "URL di base di Kokoro",
+ "Kokoro API Key": "Chiave API Kokoro (facoltativa; [Configurazione](https://github.com/remsky/Kokoro-FastAPI))",
+ "Kokoro Model": "Modello Kokoro",
+ "Kokoro Voices": "Voci Kokoro",
+ "Kokoro Voices Placeholder": "Lascia vuoto per caricare le voci del server, oppure inserisci af_heart, bf_emma, hf_alpha",
+ "Kokoro Voices Unavailable": "Impossibile caricare le voci Kokoro. La selezione precedente viene mantenuta. Controlla la connessione o inserisci le voci manualmente.",
+ "Kokoro Base URL Placeholder": "http://localhost:8880/v1",
"Login Required": "Accesso richiesto",
"Please login to access settings": "Effettua l'accesso per accedere alle impostazioni",
"Username": "Nome utente",
diff --git a/webui/i18n/ko.json b/webui/i18n/ko.json
index e93ac682..071c3119 100644
--- a/webui/i18n/ko.json
+++ b/webui/i18n/ko.json
@@ -1,6 +1,13 @@
{
"Language": "한국어",
"Translation": {
+ "Kokoro Base URL": "Kokoro 기본 URL",
+ "Kokoro API Key": "Kokoro API 키 (선택 사항; [설정 안내](https://github.com/remsky/Kokoro-FastAPI))",
+ "Kokoro Model": "Kokoro 모델",
+ "Kokoro Voices": "Kokoro 음성",
+ "Kokoro Voices Placeholder": "비워 두면 서버의 음성을 불러옵니다. 직접 입력 예: af_heart, bf_emma, hf_alpha",
+ "Kokoro Voices Unavailable": "Kokoro 음성 목록을 불러오지 못했습니다. 기존 선택은 유지됩니다. 연결을 확인하거나 음성을 직접 입력하세요.",
+ "Kokoro Base URL Placeholder": "http://localhost:8880/v1",
"Login Required": "로그인 필요",
"Please login to access settings": "설정에 접근하려면 로그인하세요",
"Username": "사용자 이름",
diff --git a/webui/i18n/pt.json b/webui/i18n/pt.json
index 41d55170..b4436967 100644
--- a/webui/i18n/pt.json
+++ b/webui/i18n/pt.json
@@ -1,6 +1,13 @@
{
"Language": "Português",
"Translation": {
+ "Kokoro Base URL": "URL base do Kokoro",
+ "Kokoro API Key": "Chave de API do Kokoro (opcional; [Configuração](https://github.com/remsky/Kokoro-FastAPI))",
+ "Kokoro Model": "Modelo do Kokoro",
+ "Kokoro Voices": "Vozes do Kokoro",
+ "Kokoro Voices Placeholder": "Deixe vazio para carregar as vozes do servidor ou digite af_heart, bf_emma, hf_alpha",
+ "Kokoro Voices Unavailable": "Não foi possível carregar as vozes do Kokoro. A seleção anterior foi mantida. Verifique a conexão ou informe as vozes manualmente.",
+ "Kokoro Base URL Placeholder": "http://localhost:8880/v1",
"Login Required": "Login Necessário",
"Please login to access settings": "Por favor, faça login para acessar as configurações",
"Username": "Nome de usuário",
diff --git a/webui/i18n/ru.json b/webui/i18n/ru.json
index 14070b61..488b1f30 100644
--- a/webui/i18n/ru.json
+++ b/webui/i18n/ru.json
@@ -1,6 +1,13 @@
{
"Language": "Русский",
"Translation": {
+ "Kokoro Base URL": "Базовый URL Kokoro",
+ "Kokoro API Key": "API-ключ Kokoro (необязательно; [Настройка](https://github.com/remsky/Kokoro-FastAPI))",
+ "Kokoro Model": "Модель Kokoro",
+ "Kokoro Voices": "Голоса Kokoro",
+ "Kokoro Voices Placeholder": "Оставьте пустым для загрузки голосов с сервера или укажите af_heart, bf_emma, hf_alpha",
+ "Kokoro Voices Unavailable": "Не удалось загрузить голоса Kokoro. Предыдущий выбор сохранён. Проверьте подключение или укажите голоса вручную.",
+ "Kokoro Base URL Placeholder": "http://localhost:8880/v1",
"Login Required": "Требуется вход",
"Please login to access settings": "Войдите, чтобы открыть настройки",
"Username": "Имя пользователя",
diff --git a/webui/i18n/tr.json b/webui/i18n/tr.json
index f08c6f6e..08d97507 100644
--- a/webui/i18n/tr.json
+++ b/webui/i18n/tr.json
@@ -1,6 +1,13 @@
{
"Language": "Türkçe",
"Translation": {
+ "Kokoro Base URL": "Kokoro Temel URL",
+ "Kokoro API Key": "Kokoro API Anahtarı (isteğe bağlı; [Kurulum](https://github.com/remsky/Kokoro-FastAPI))",
+ "Kokoro Model": "Kokoro Modeli",
+ "Kokoro Voices": "Kokoro Sesleri",
+ "Kokoro Voices Placeholder": "Sunucu seslerini yüklemek için boş bırakın veya af_heart, bf_emma, hf_alpha girin",
+ "Kokoro Voices Unavailable": "Kokoro sesleri yüklenemedi. Önceki seçim korundu. Bağlantıyı kontrol edin veya sesleri elle girin.",
+ "Kokoro Base URL Placeholder": "http://localhost:8880/v1",
"Login Required": "Giriş Gerekli",
"Please login to access settings": "Ayarlara erişmek için lütfen giriş yapın",
"Username": "Kullanıcı Adı",
diff --git a/webui/i18n/vi.json b/webui/i18n/vi.json
index 8c5b30ce..b15f239b 100644
--- a/webui/i18n/vi.json
+++ b/webui/i18n/vi.json
@@ -1,6 +1,13 @@
{
"Language": "Tiếng Việt",
"Translation": {
+ "Kokoro Base URL": "URL cơ sở Kokoro",
+ "Kokoro API Key": "Khóa API Kokoro (không bắt buộc; [Hướng dẫn cấu hình](https://github.com/remsky/Kokoro-FastAPI))",
+ "Kokoro Model": "Mô hình Kokoro",
+ "Kokoro Voices": "Giọng Kokoro",
+ "Kokoro Voices Placeholder": "Để trống để tải giọng từ máy chủ, hoặc nhập af_heart, bf_emma, hf_alpha",
+ "Kokoro Voices Unavailable": "Không thể tải danh sách giọng Kokoro. Lựa chọn trước đó được giữ lại. Kiểm tra kết nối hoặc nhập giọng thủ công.",
+ "Kokoro Base URL Placeholder": "http://localhost:8880/v1",
"Login Required": "Yêu cầu đăng nhập",
"Please login to access settings": "Vui lòng đăng nhập để truy cập cài đặt",
"Username": "Tên đăng nhập",
diff --git a/webui/i18n/zh.json b/webui/i18n/zh.json
index 67e1045a..1f4c1440 100644
--- a/webui/i18n/zh.json
+++ b/webui/i18n/zh.json
@@ -1,6 +1,7 @@
{
"Language": "简体中文",
"Translation": {
+ "Kokoro Voices Unavailable": "暂时无法获取 Kokoro 音色,已保留原有选择。请检查服务连接,或手动填写音色。",
"Login Required": "需要登录",
"Please login to access settings": "请登录后访问配置设置 (:gray[默认用户名: admin, 密码: admin, 您可以在 config.toml 中修改])",
"Username": "用户名",