From 2d4492399dcdae4c967707fe3f8313fe3b32a0e7 Mon Sep 17 00:00:00 2001
From: Sirshendu Ganguly <sirshendu.ganguly2015@gmail.com>
Date: Mon, 20 Apr 2026 15:15:04 +0530
Subject: [PATCH 1/3] Added support for voices in audioInference

---
 runware/types.py | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/runware/types.py b/runware/types.py
index 3f1dc14..15c9390 100644
--- a/runware/types.py
+++ b/runware/types.py
@@ -1813,10 +1813,17 @@ def request_key(self) -> str:
         return "inputs"
 
 
+@dataclass
+class IAudioVoice(SerializableMixin):
+    speaker: str
+    voice: str
+
+
 @dataclass
 class IAudioSpeech(SerializableMixin):
     text: Optional[str] = None  
     voice: Optional[str] = None
+    voices: Optional[List[Union[IAudioVoice, Dict[str, Any]]]] = None
     language: Optional[str] = None
     speed: Optional[float] = None
     volume: Optional[int] = None
@@ -1824,6 +1831,13 @@ class IAudioSpeech(SerializableMixin):
     emotion: Optional[str] = None
     tone: Optional[List[str]] = None  
 
+    def __post_init__(self):
+        if self.voices is not None:
+            self.voices = [
+                IAudioVoice(**v) if isinstance(v, dict) else v
+                for v in self.voices
+            ]
+
     @property
     def request_key(self) -> str:
         return "speech"

From e3ff029b3055915fb514a214c364f8d2fca046a1 Mon Sep 17 00:00:00 2001
From: Sirshendu Ganguly <sirshendu.ganguly2015@gmail.com>
Date: Mon, 20 Apr 2026 15:26:39 +0530
Subject: [PATCH 2/3] Fixed post_init

---
 runware/types.py | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/runware/types.py b/runware/types.py
index 15c9390..db45361 100644
--- a/runware/types.py
+++ b/runware/types.py
@@ -1833,10 +1833,14 @@ class IAudioSpeech(SerializableMixin):
 
     def __post_init__(self):
         if self.voices is not None:
-            self.voices = [
-                IAudioVoice(**v) if isinstance(v, dict) else v
-                for v in self.voices
-            ]
+            normalized_voices = []
+            for v in self.voices:
+                if isinstance(v, dict):
+                    normalized_voices.append(IAudioVoice(**v))
+                else:
+                    normalized_voices.append(v)
+
+            self.voices = normalized_voices
 
     @property
     def request_key(self) -> str:

From 92a943fbd037f6af197c39730415d417396e401c Mon Sep 17 00:00:00 2001
From: Sirshendu Ganguly <sirshendu.ganguly2015@gmail.com>
Date: Mon, 20 Apr 2026 15:39:57 +0530
Subject: [PATCH 3/3] Fixed validations

---
 runware/types.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/runware/types.py b/runware/types.py
index db45361..80517d9 100644
--- a/runware/types.py
+++ b/runware/types.py
@@ -1832,7 +1832,7 @@ class IAudioSpeech(SerializableMixin):
     tone: Optional[List[str]] = None  
 
     def __post_init__(self):
-        if self.voices is not None:
+        if self.voices is not None and isinstance(self.voices, (list, tuple)):
             normalized_voices = []
             for v in self.voices:
                 if isinstance(v, dict):