Merge pull request #107 from Runware/feature-addAudioModelsXai/Dia

Added workflow for xAI TTS and Dia TTS

__init__.py

@@ -53,7 +53,6 @@ from .modules.providerSettings.viduProviderSettings import RunwareViduProviderSe
 from .modules.mediaUpload import runwareMediaUpload
 from .modules.inputAudios import inputAudios
 from .modules.audioInput import RunwareAudioInput
-from .modules.audioInputCombine import RunwareAudioInputCombine
 from .modules.speechInput import RunwareSpeechInput
 from .modules.speechInputCombine import RunwareSpeechInputCombine
 from .modules.referenceVideos import referenceVideos
@@ -163,7 +162,6 @@ NODE_CLASS_MAPPINGS = {
     "Runware Media Upload": runwareMediaUpload,
     "Runware Input Audios": inputAudios,
     "Runware Video Audio Input": RunwareAudioInput,
-    "Runware Audio Inputs Combine": RunwareAudioInputCombine,
     "Runware Video Speech Input": RunwareSpeechInput,
     "Runware Speech Inputs Combine": RunwareSpeechInputCombine,
     "Runware Reference Videos": referenceVideos,

clientlibs/utils.js

@@ -1212,6 +1212,8 @@ function videoSettingsToggleHandler(settingsNode) {
 function audioInferenceSpeechToggleHandler(speechNode) {
     if (!speechNode?.widgets) return;
 
+    const useVoiceWidget = speechNode.widgets.find(w => w && w.name === "useVoice");
+    const voiceWidget = speechNode.widgets.find(w => w && w.name === "voice");
     const useSpeedWidget = speechNode.widgets.find(w => w && w.name === "useSpeed");
     const speedWidget = speechNode.widgets.find(w => w && w.name === "speed");
     const useVolumeWidget = speechNode.widgets.find(w => w && w.name === "useVolume");
@@ -1244,6 +1246,7 @@ function audioInferenceSpeechToggleHandler(speechNode) {
     if (usePitchWidget && pitchWidget) toggleWidgetState(usePitchWidget, pitchWidget, "pitch");
     if (useEmotionWidget && emotionWidget) toggleWidgetState(useEmotionWidget, emotionWidget, "emotion");
     if (useToneWidget && toneWidget) toggleWidgetState(useToneWidget, toneWidget, "tone");
+    if (useVoiceWidget && voiceWidget) toggleWidgetState(useVoiceWidget, voiceWidget, "voice");
 }
 
 function acceleratorOptionsToggleHandler(acceleratorNode) {
@@ -2763,6 +2766,12 @@ function audioModelSearchFilterHandler(audioModelSearchNode) {
         "Ace": [
             "runware:ace-step@0 (ACE Step v1 3.5B)",
         ],
+        "Dia": [
+            "runware:dia@v1.0 (Dia 1.6B)",
+        ],
+        "xAI": [
+            "xai:voice@tts (xAI TTS)",
+        ],
         "MiniMax": [
             "minimax:speech@2.8 (MiniMax Speech 2.8)",
         ],

modules/audioInferenceSpeech.py

@@ -26,12 +26,16 @@ class RunwareAudioInferenceSpeech:
                     "default": "Hello, this is a test of the speech synthesis model.",
                     "tooltip": "Text to synthesize as speech (speech.text)",
                 }),
-                "voice": ("STRING", {
-                    "default": "English_Whispering_girl",
-                    "tooltip": "Voice identifier - see System Voice ID List (speech.voice)",
-                }),
             },
             "optional": {
+                "useVoice": ("BOOLEAN", {
+                    "default": False,
+                    "tooltip": "Enable to include voice in speech synthesis (speech.voice).",
+                }),
+                "voice": ("STRING", {
+                    "default": "",
+                    "tooltip": "Voice identifier.",
+                }),
                 "useSpeed": ("BOOLEAN", {"default": True, "tooltip": "Include speed (speech.speed)"}),
                 "speed": ("FLOAT", {
                     "default": 1.0,
@@ -78,6 +82,7 @@ class RunwareAudioInferenceSpeech:
         """Build the speech object for the API (connect to Runware Audio Inference 'speech' input)."""
         text = (kwargs.get("text") or "").strip()
         voice = (kwargs.get("voice") or "").strip()
+        use_voice = kwargs.get("useVoice", False)
         use_speed = kwargs.get("useSpeed", True)
         speed = float(kwargs.get("speed", 1.0))
         use_volume = kwargs.get("useVolume", True)
@@ -89,7 +94,9 @@ class RunwareAudioInferenceSpeech:
         use_tone = kwargs.get("useTone", False)
         tone_str = kwargs.get("tone", "") or ""
 
-        speech: Dict[str, Any] = {"text": text, "voice": voice}
+        speech: Dict[str, Any] = {"text": text}
+        if use_voice:
+            speech["voice"] = voice
         if use_speed:
             speech["speed"] = speed
         if use_volume:
@@ -103,6 +110,8 @@ class RunwareAudioInferenceSpeech:
             if tone_lines:
                 speech["tone"] = tone_lines
 
+        if not text:
+            return ({},)
         return (speech,)
 
 

modules/audioInputCombine.py

@@ -1,60 +0,0 @@
-"""
-Runware Audio Input Combine Node
-Combines multiple audio inputs into a list
-"""
-
-from typing import List, Dict, Any
-
-
-class RunwareAudioInputCombine:
-    """Runware Audio Input Combine Node"""
-
-    @classmethod
-    def INPUT_TYPES(cls):
-        return {
-            "required": {
-                "Audio Input 1": ("RUNWAREAUDIOINPUT", {
-                    "tooltip": "Connect a Runware Audio Input node.",
-                }),
-            },
-            "optional": {
-                "Audio Input 2": ("RUNWAREAUDIOINPUT", {
-                    "tooltip": "Connect a Runware Audio Input node.",
-                }),
-                "Audio Input 3": ("RUNWAREAUDIOINPUT", {
-                    "tooltip": "Connect a Runware Audio Input node.",
-                }),
-                "Audio Input 4": ("RUNWAREAUDIOINPUT", {
-                    "tooltip": "Connect a Runware Audio Input node.",
-                }),
-            },
-        }
-
-    DESCRIPTION = "Combine multiple audio inputs to connect with Runware Video Inference Inputs."
-    FUNCTION = "audioInputCombine"
-    RETURN_TYPES = ("RUNWAREAUDIOINPUT",)
-    RETURN_NAMES = ("Audio Inputs",)
-    CATEGORY = "Runware"
-
-    def audioInputCombine(self, **kwargs):
-        """Combine audio inputs into a list"""
-        audioInputs = []
-        
-        for i in range(1, 5):
-            audioKey = f"Audio Input {i}"
-            audioInput = kwargs.get(audioKey)
-            if audioInput is not None and isinstance(audioInput, dict) and len(audioInput) > 0:
-                audioInputs.append(audioInput)
-        
-        return (audioInputs,)
-
-
-# Node class mappings
-NODE_CLASS_MAPPINGS = {
-    "RunwareAudioInputCombine": RunwareAudioInputCombine,
-}
-
-NODE_DISPLAY_NAME_MAPPINGS = {
-    "RunwareAudioInputCombine": "Runware Audio Inputs Combine",
-}
-

modules/audioModelSearch.py

@@ -17,6 +17,12 @@ class RunwareAudioModelSearch:
         "Ace": [
             "runware:ace-step@0 (ACE Step v1 3.5B)",
         ],
+        "Dia": [
+            "runware:dia@v1.0 (Dia 1.6B)",
+        ],
+        "xAI": [
+            "xai:voice@tts (xAI TTS)",
+        ],
         "MiniMax": [
             "minimax:speech@2.8 (MiniMax Speech 2.8)",
         ],
@@ -32,6 +38,8 @@ class RunwareAudioModelSearch:
         "KlingAI",
         "Mirelo",
         "Ace",
+        "Dia",
+        "xAI",
         "MiniMax",
         "Inworld",
     ]

modules/videoInferenceInputs.py

@@ -24,7 +24,7 @@ class videoInferenceInputs:
                     "tooltip": "Connect the mediaUUID output from Runware Media Upload node with audio file. This audio will be synchronized with the generated video. (Legacy: single audio string)"
                 }),
                 "Audio Inputs": ("RUNWAREAUDIOINPUT", {
-                    "tooltip": "Connect Runware Video Audio Input node or Runware Audio Inputs Combine node to provide audio inputs with reference IDs."
+                    "tooltip": "Connect Runware Video Audio Input node(s) to provide audio inputs with reference IDs."
                 }),
                 "Speech Inputs": ("RUNWARESPEECHINPUT", {
                     "tooltip": "Connect Runware Video Speech Input node or Runware Speech Inputs Combine node to provide TTS speech inputs with reference IDs."