Add voice state tags, SuperTTS configs, and demo WS log groups.

Parse leading <state> tags from LLM replies and emit response.state over the product websocket while stripping tags from TTS/text streams. Add FastGPT+Xfyun voice configs (including state-enabled preset), SuperTTS support, and context sync for interrupted turns. Refresh the voice demo with a state indicator and collapsible audio delta websocket log groups.

Co-authored-by: Cursor <cursoragent@cursor.com>

src/voice/services.py

@@ -10,11 +10,13 @@ from pipecat.services.openai._constants import OPENAI_SAMPLE_RATE
 from pipecat.services.openai.llm import OpenAILLMService
 from pipecat.services.openai.stt import OpenAISTTService
 from pipecat.services.openai.tts import VALID_VOICES, OpenAITTSService
+from pipecat.services.tts_service import TextAggregationMode
 from pipecat.transcriptions.language import Language
 
 from .config import AudioConfig, LLMConfig, STTConfig, TTSConfig
 from .fastgpt_llm import FastGPTLLMService, FastGPTLLMSettings
 from .xfyun_asr import DEFAULT_XFYUN_ASR_URL, XfyunASRService
+from .xfyun_super_tts import DEFAULT_XFYUN_SUPER_TTS_URL, XfyunSuperTTSService
 from .xfyun_tts import DEFAULT_XFYUN_TTS_URL, XfyunTTSService
 
 
@@ -54,12 +56,13 @@ def create_llm_service(
     session_variables: dict | None = None,
     greeting_prompt: str | None = None,
 ):
-    if config.provider == "fastgpt":
+    if config.is_fastgpt:
         variables = {**config.variables, **(session_variables or {})}
         return FastGPTLLMService(
             api_key=config.api_key,
             base_url=config.base_url or "http://localhost:3000",
             chat_id=chat_id or config.chat_id,
+            app_id=config.app_id,
             send_system_prompt=config.send_system_prompt,
             greeting_prompt=greeting_prompt,
             timeout=config.timeout_sec,
@@ -70,7 +73,11 @@ def create_llm_service(
             ),
         )
 
-    _require_provider(config.provider, "openai", "llm")
+    if not config.is_openai:
+        supported = ", ".join(sorted(("openai", "fastgpt", "llm")))
+        raise ValueError(
+            f"Unsupported llm provider {config.provider!r}; expected one of: {supported}"
+        )
     return OpenAILLMService(
         api_key=config.api_key or None,
         base_url=config.base_url,
@@ -102,6 +109,30 @@ def create_tts_service(config: TTSConfig, audio: AudioConfig):
             timeout=config.timeout_sec,
         )
 
+    if config.provider in ("xfyun_super", "xfyun_super_tts"):
+        source_sample_rate = config.source_sample_rate_hz or 24000
+        if source_sample_rate not in (8000, 16000, 24000):
+            raise ValueError(
+                "Xfyun Super TTS source_sample_rate_hz must be 8000, 16000, or 24000"
+            )
+        text_aggregation_mode = config.text_aggregation_mode or TextAggregationMode.TOKEN
+        return XfyunSuperTTSService(
+            app_id=config.app_id,
+            api_key=config.api_key or "",
+            api_secret=config.api_secret,
+            voice=config.voice,
+            url=config.base_url or DEFAULT_XFYUN_SUPER_TTS_URL,
+            sample_rate=audio.sample_rate_hz,
+            source_sample_rate=source_sample_rate,
+            encoding=config.aue,
+            speed=config.speed,
+            volume=config.volume,
+            pitch=config.pitch,
+            oral_level=config.oral_level,
+            text_aggregation_mode=text_aggregation_mode,
+            open_timeout=config.timeout_sec,
+        )
+
     _require_provider(config.provider, "openai", "tts")
     service_class = OpenAITTSService if config.voice in VALID_VOICES else OpenAICompatibleTTSService
     return service_class(