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.

Add Chatterbox WebUI settings, provider tests, and clarify voice_volume

Paulo C. Benetti815e9ce30492
diff --git a/app/services/voice.py b/app/services/voice.py
index 190fa546..1d8072f0 100644
--- a/app/services/voice.py
+++ b/app/services/voice.py
@@ -1359,6 +1359,10 @@ def chatterbox_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 for parity with the other TTS providers but is
+ # intentionally not sent: the OpenAI /audio/speech contract has no volume
+ # field, so Chatterbox servers ignore it. Adjust loudness via voice_rate
+ # (speed) or in post-processing instead.
for i in range(3):
try:
diff --git a/test/services/test_voice.py b/test/services/test_voice.py
index c4b5f58c..57128114 100644
--- a/test/services/test_voice.py
+++ b/test/services/test_voice.py
@@ -504,6 +504,127 @@ class TestVoiceService(unittest.TestCase):
self.assertEqual(getattr(sub_maker, "subs", []), ["小米语音合成测试", "第二句话"])
self.assertEqual(len(getattr(sub_maker, "offset", [])), 2)
+ def test_chatterbox_voice_helpers(self):
+ """is_chatterbox_voice / get_chatterbox_voices basics and normalisation."""
+ self.assertTrue(vs.is_chatterbox_voice("chatterbox:default-Female"))
+ self.assertFalse(vs.is_chatterbox_voice("elevenlabs:abc:Rachel"))
+ self.assertFalse(vs.is_chatterbox_voice(""))
+ self.assertFalse(vs.is_chatterbox_voice(None))
+
+ # list entries are normalised to the chatterbox:<name> dispatcher format,
+ # and entries that are already prefixed are left untouched
+ with patch.object(
+ vs.config,
+ "chatterbox",
+ {"voices": ["narrator-Male", "chatterbox:host"]},
+ ):
+ self.assertEqual(
+ vs.get_chatterbox_voices(),
+ ["chatterbox:narrator-Male", "chatterbox:host"],
+ )
+
+ # a comma-separated string is also accepted (TOML-friendly)
+ with patch.object(vs.config, "chatterbox", {"voices": "alpha, beta ,"}):
+ self.assertEqual(
+ vs.get_chatterbox_voices(),
+ ["chatterbox:alpha", "chatterbox:beta"],
+ )
+
+ # with nothing configured the dropdown still gets a usable default
+ with patch.object(vs.config, "chatterbox", {}):
+ self.assertEqual(vs.get_chatterbox_voices(), ["chatterbox:default-Female"])
+
+ def test_chatterbox_tts_posts_to_openai_compatible_endpoint(self):
+ """Success path: POST /audio/speech, write audio, return legacy SubMaker."""
+
+ class _FakeResponse:
+ status_code = 200
+ content = b"RIFF-fake-wav"
+ text = ""
+
+ class _FakeClip:
+ duration = 3.5
+
+ def close(self):
+ pass
+
+ captured = {}
+
+ def _fake_post(url, json=None, headers=None, timeout=None):
+ captured["url"] = url
+ captured["json"] = json
+ captured["headers"] = headers
+ return _FakeResponse()
+
+ with tempfile.TemporaryDirectory() as tmp_dir, patch.object(
+ vs.config,
+ "chatterbox",
+ {
+ "base_url": "http://localhost:4123/v1/",
+ "api_key": "secret",
+ "model_id": "chatterbox",
+ },
+ ), patch.object(
+ vs.requests, "post", side_effect=_fake_post
+ ) as post, patch.object(
+ vs, "AudioFileClip", return_value=_FakeClip()
+ ):
+ voice_file = str(Path(tmp_dir) / "chatterbox.mp3")
+ sub_maker = vs.chatterbox_tts(
+ text="Hello world. Second sentence.",
+ voice="default",
+ voice_file=voice_file,
+ voice_rate=1.2,
+ voice_volume=1.0,
+ )
+ generated_audio = Path(voice_file).read_bytes()
+
+ post.assert_called_once()
+ # trailing slash on base_url is stripped before appending /audio/speech
+ self.assertEqual(captured["url"], "http://localhost:4123/v1/audio/speech")
+ self.assertEqual(captured["json"]["model"], "chatterbox")
+ self.assertEqual(captured["json"]["voice"], "default")
+ self.assertEqual(captured["json"]["input"], "Hello world. Second sentence.")
+ self.assertAlmostEqual(captured["json"]["speed"], 1.2)
+ # api_key is forwarded as a bearer token
+ self.assertEqual(captured["headers"].get("Authorization"), "Bearer secret")
+ # volume is intentionally not part of the OpenAI speech payload
+ self.assertNotIn("volume", captured["json"])
+ self.assertEqual(generated_audio, b"RIFF-fake-wav")
+ self.assertIsNotNone(sub_maker)
+ self.assertTrue(getattr(sub_maker, "subs", []))
+
+ def test_chatterbox_tts_requires_base_url(self):
+ """Missing base_url short-circuits without any network call."""
+ with patch.object(
+ vs.config, "chatterbox", {"base_url": ""}
+ ), patch.object(vs.requests, "post") as post:
+ result = vs.chatterbox_tts(
+ text="hi", voice="default", voice_file="unused.mp3"
+ )
+ self.assertIsNone(result)
+ post.assert_not_called()
+
+ def test_chatterbox_tts_returns_none_on_http_error(self):
+ """A non-200 response is retried up to 3 times, then fails to None."""
+
+ class _FakeResponse:
+ status_code = 500
+ content = b""
+ text = "boom"
+
+ with tempfile.TemporaryDirectory() as tmp_dir, patch.object(
+ vs.config, "chatterbox", {"base_url": "http://localhost:4123/v1"}
+ ), patch.object(
+ vs.requests, "post", return_value=_FakeResponse()
+ ) as post:
+ voice_file = str(Path(tmp_dir) / "chatterbox.mp3")
+ result = vs.chatterbox_tts(
+ text="hi", voice="default", voice_file=voice_file
+ )
+ self.assertIsNone(result)
+ self.assertEqual(post.call_count, 3)
+
def test_generate_subtitle_keeps_edge_provider_for_gemini_legacy_submaker(self):
"""
验证 Gemini TTS 返回的 legacy 字幕结构在 edge provider 下可以直接产出
diff --git a/webui/Main.py b/webui/Main.py
index cddc26d1..850c8a32 100644
--- a/webui/Main.py
+++ b/webui/Main.py
@@ -1259,6 +1259,58 @@ with middle_panel:
config.elevenlabs["api_key"] = elevenlabs_api_key
+ # Chatterbox API settings section (self-hosted, OpenAI-compatible)
+ if selected_tts_server == "chatterbox" or (
+ voice_name and voice.is_chatterbox_voice(voice_name)
+ ):
+ chatterbox_base_url = st.text_input(
+ tr("Chatterbox Base URL"),
+ value=config.chatterbox.get("base_url", ""),
+ key="chatterbox_base_url_input",
+ placeholder="http://localhost:4123/v1",
+ )
+ config.chatterbox["base_url"] = (chatterbox_base_url or "").strip()
+
+ chatterbox_api_key = st.text_input(
+ tr("Chatterbox API Key"),
+ value=config.chatterbox.get("api_key", ""),
+ type="password",
+ key="chatterbox_api_key_input",
+ )
+ config.chatterbox["api_key"] = chatterbox_api_key
+
+ chatterbox_model = st.text_input(
+ tr("Chatterbox Model"),
+ value=config.chatterbox.get("model_id", "chatterbox") or "chatterbox",
+ key="chatterbox_model_input",
+ )
+ config.chatterbox["model_id"] = (chatterbox_model or "chatterbox").strip()
+
+ _saved_chatterbox_voices = config.chatterbox.get("voices", [])
+ if isinstance(_saved_chatterbox_voices, list):
+ _saved_chatterbox_voices = ", ".join(_saved_chatterbox_voices)
+ chatterbox_voices = st.text_input(
+ tr("Chatterbox Voices"),
+ value=str(_saved_chatterbox_voices or ""),
+ key="chatterbox_voices_input",
+ placeholder="default-Female, narrator-Male",
+ )
+ config.chatterbox["voices"] = [
+ v.strip() for v in chatterbox_voices.split(",") if v.strip()
+ ]
+
+ st.info(
+ "Chatterbox TTS Settings (self-hosted):\n"
+ "- Run an OpenAI-compatible Chatterbox server (e.g. "
+ "devnen/Chatterbox-TTS-Server or travisvn/chatterbox-tts-api) and "
+ "set Base URL to its /v1 endpoint\n"
+ "- Voices is a comma-separated list of voice names your server "
+ "exposes; add a -Female or -Male suffix only to label the gender "
+ "in this dropdown\n"
+ "- Speech Volume is not applied for Chatterbox (the OpenAI "
+ "/audio/speech API has no volume field); use Speech Rate instead"
+ )
+
params.voice_volume = st.selectbox(
tr("Speech Volume"),
options=[0.6, 0.8, 1.0, 1.2, 1.5, 2.0, 3.0, 4.0, 5.0],
diff --git a/webui/i18n/en.json b/webui/i18n/en.json
index e598b946..3732f87d 100644
--- a/webui/i18n/en.json
+++ b/webui/i18n/en.json
@@ -126,6 +126,10 @@
"MiMo TTS Settings": "Xiaomi MiMo TTS Settings",
"ElevenLabs API Key": "ElevenLabs API Key",
"ElevenLabs Model": "ElevenLabs Model",
+ "Chatterbox Base URL": "Chatterbox Base URL",
+ "Chatterbox API Key": "Chatterbox API Key",
+ "Chatterbox Model": "Chatterbox Model",
+ "Chatterbox Voices": "Chatterbox Voices",
"Uses Xiaomi MiMo V2.5 TTS preset voices": "Uses Xiaomi MiMo V2.5 TTS preset voices",
"Speed and volume are currently handled by the provider defaults": "Speed and volume are currently handled by the provider defaults",
"Hide Log": "Hide Log",
diff --git a/webui/i18n/ru.json b/webui/i18n/ru.json
index f6ecb626..9a2ce862 100644
--- a/webui/i18n/ru.json
+++ b/webui/i18n/ru.json
@@ -122,6 +122,10 @@
"MiMo TTS Settings": "Настройки TTS Xiaomi MiMo",
"ElevenLabs API Key": "API-ключ ElevenLabs",
"ElevenLabs Model": "Модель ElevenLabs",
+ "Chatterbox Base URL": "Базовый URL Chatterbox",
+ "Chatterbox API Key": "API-ключ Chatterbox",
+ "Chatterbox Model": "Модель Chatterbox",
+ "Chatterbox Voices": "Голоса Chatterbox",
"Uses Xiaomi MiMo V2.5 TTS preset voices": "Используются предустановленные голоса Xiaomi MiMo V2.5",
"Speed and volume are currently handled by the provider defaults": "Скорость и громкость сейчас задаются настройками провайдера по умолчанию",
"Hide Log": "Скрыть журнал",