Fix Piper TTS voices stuck on "Loading voices..."

requirements.txt

@@ -1,5 +1,6 @@
 customtkinter
 edge-tts>=7.2.3
+piper-tts>=1.2.0
 langid>=1.1.6
 sounddevice
 soundfile>=0.12.0

src/config/settings_manager.py

@@ -118,7 +118,10 @@ class SettingsManager:
         "visible_buttons": ["speak", "stop", "clear", "voice", "overlay"],  # Toggleable buttons to show (settings is always visible)
 
         # Overlay Settings
-        "overlay_visible": True  # Whether the recording overlay is visible
+        "overlay_visible": True,  # Whether the recording overlay is visible
+
+        # TTS Provider Selection
+        "tts_provider": "edge",  # "edge" (Edge TTS, online) or "piper" (Piper TTS, offline)
     }
 
 
@@ -315,6 +318,11 @@ def set(self, key, value):
                 logger.warning("Invalid appearance mode: %s, using default", value)
                 value = self.DEFAULT_SETTINGS["appearance_mode"]
 
+            # Validate TTS provider
+            if key == "tts_provider" and value not in ["edge", "piper"]:
+                logger.warning("Invalid TTS provider: %s, using default", value)
+                value = self.DEFAULT_SETTINGS["tts_provider"]
+
             self._settings[key] = value
 
     def get_all(self):
@@ -378,6 +386,11 @@ def validate_settings(self) -> list:
             if appearance not in ["Dark", "Light", "System"]:
                 issues.append(f"Invalid appearance mode: {appearance}")
 
+            # Check TTS provider
+            tts_provider = self._settings.get("tts_provider")
+            if tts_provider not in ["edge", "piper"]:
+                issues.append(f"Invalid TTS provider: {tts_provider}")
+
             # Check numeric range settings
             rate = self._settings.get("rate")
             if not isinstance(rate, (int, float)) or not (-100 <= rate <= 100):

src/gui/settings_tabs/__init__.py

@@ -34,6 +34,7 @@
 from .behavior_tab import BehaviorTab
 from .vrchat_osc_tab import VRChatOSCTab
 from .advanced_tab import AdvancedTab
+from .tts_provider_tab import TTSProviderTab
 
 __all__ = [
     'BaseTab',
@@ -45,4 +46,5 @@
     'BehaviorTab',
     'VRChatOSCTab',
     'AdvancedTab',
+    'TTSProviderTab',
 ]

src/gui/settings_tabs/tts_provider_tab.py

@@ -0,0 +1,148 @@
+"""
+TTS Provider Tab
+Settings for choosing between available TTS providers (online / offline).
+"""
+import customtkinter as ctk
+from typing import Any, List, Dict
+
+from .base_tab import BaseTab
+from ..theme_constants import (
+    FONT_SM, FONT_MD, FONT_LG, FONT_XL, FONT_WEIGHT_BOLD,
+    COLOR_INFO,
+)
+
+
+_PROVIDER_OPTIONS = [
+    "edge",
+    "piper",
+]
+
+_PROVIDER_LABELS = {
+    "edge": "Edge TTS (Online)",
+    "piper": "Piper TTS (Offline)",
+}
+
+_PROVIDER_DESCRIPTIONS = {
+    "edge": (
+        "Microsoft Edge Text-to-Speech\n\n"
+        "• Requires an active internet connection\n"
+        "• 100+ high-quality neural voices\n"
+        "• Broad language and dialect support\n"
+        "• Rate, volume, and pitch controls"
+    ),
+    "piper": (
+        "Piper — Open-Source Offline Neural TTS\n\n"
+        "• Works fully offline — no internet needed\n"
+        "• Privacy-focused (no data sent externally)\n"
+        "• Open-source voice models (ONNX-based)\n"
+        "• Voice models are downloaded on first use\n"
+        "• Rate and volume controls supported\n"
+        "• Note: pitch control is not supported by Piper"
+    ),
+}
+
+
+class TTSProviderTab(BaseTab):
+    """Tab for selecting the active TTS provider."""
+
+    def _create_content(self):
+        """Build the TTS Provider tab UI."""
+        self.setup_layout()
+
+        # Title
+        ctk.CTkLabel(
+            self.scroll,
+            text="TTS Provider",
+            font=ctk.CTkFont(size=FONT_XL, weight=FONT_WEIGHT_BOLD),
+        ).pack(anchor="w", pady=(10, 5))
+
+        intro_label = ctk.CTkLabel(
+            self.scroll,
+            text=(
+                "Choose which text-to-speech engine CriTTS uses. "
+                "Changes take effect after saving and will also update the available voice list."
+            ),
+            font=ctk.CTkFont(size=FONT_SM),
+            text_color="gray",
+            wraplength=550,
+        )
+        intro_label.pack(anchor="w", pady=(0, 15))
+        self.add_wraplength_label(intro_label)
+
+        # --- Provider selection ---
+        self.create_section_header("Active Provider").pack(anchor="w", pady=(5, 5))
+
+        current = self.settings.get("tts_provider", "edge")
+
+        self._provider_var = ctk.StringVar(value=_PROVIDER_LABELS.get(current, _PROVIDER_LABELS["edge"]))
+
+        self._provider_dropdown = ctk.CTkComboBox(
+            self.scroll,
+            variable=self._provider_var,
+            values=[_PROVIDER_LABELS[k] for k in _PROVIDER_OPTIONS],
+            font=ctk.CTkFont(size=FONT_MD),
+            state="readonly",
+            width=320,
+            command=self._on_provider_changed,
+        )
+        self._provider_dropdown.pack(anchor="w", pady=(5, 10))
+
+        self.create_separator(self.scroll).pack(fill="x", pady=(5, 10))
+
+        # --- Description box ---
+        self.create_section_header("Provider Details").pack(anchor="w", pady=(5, 5))
+
+        self._desc_label = ctk.CTkLabel(
+            self.scroll,
+            text=_PROVIDER_DESCRIPTIONS.get(current, ""),
+            font=ctk.CTkFont(size=FONT_SM),
+            justify="left",
+            wraplength=500,
+        )
+        self._desc_label.pack(anchor="w", pady=(5, 10))
+        self.add_wraplength_label(self._desc_label)
+
+        self.create_separator(self.scroll).pack(fill="x", pady=(5, 10))
+
+        # --- Informational note about voices ---
+        note_label = ctk.CTkLabel(
+            self.scroll,
+            text=(
+                "ℹ  After switching providers, save and reopen Settings to see "
+                "the updated voice list in the Voice tab."
+            ),
+            font=ctk.CTkFont(size=FONT_SM),
+            text_color=COLOR_INFO,
+            wraplength=500,
+            justify="left",
+        )
+        note_label.pack(anchor="w", pady=(0, 10))
+        self.add_wraplength_label(note_label)
+
+    # ------------------------------------------------------------------
+
+    def _on_provider_changed(self, selected_label: str):
+        """Update the description box when a new provider is selected."""
+        key = self._label_to_key(selected_label)
+        self._desc_label.configure(text=_PROVIDER_DESCRIPTIONS.get(key, ""))
+
+    @staticmethod
+    def _label_to_key(label: str) -> str:
+        """Reverse-lookup: human-readable label → internal key."""
+        for k, v in _PROVIDER_LABELS.items():
+            if v == label:
+                return k
+        return "edge"
+
+    # ------------------------------------------------------------------
+    # BaseTab interface
+    # ------------------------------------------------------------------
+
+    def get_settings(self) -> Dict[str, Any]:
+        return {"tts_provider": self._label_to_key(self._provider_var.get())}
+
+    def validate(self) -> List[str]:
+        key = self._label_to_key(self._provider_var.get())
+        if key not in _PROVIDER_OPTIONS:
+            return [f"Unknown TTS provider: '{self._provider_var.get()}'"]
+        return []

src/gui/settings_window.py

@@ -20,7 +20,7 @@
 from .keybind_manager import KeybindManager
 from .settings_tabs import (
     VoiceTab, AudioOutputTab, AppearanceTab, AbbreviationsTab,
-    KeybindsTab, BehaviorTab, VRChatOSCTab, AdvancedTab
+    KeybindsTab, BehaviorTab, VRChatOSCTab, AdvancedTab, TTSProviderTab
 )
 from .utils.scroll_utils import prevent_scroll_propagation
 from .theme_constants import (
@@ -133,7 +133,7 @@ def _create_window(self):
 
         # Voice Settings Tab
         voice_frame = self.tabview.add("Voice")
-        self.voice_tab_obj = VoiceTab(voice_frame, self.settings, self.tts_engine, self.audio_router, self._on_change_placeholder)
+        self.voice_tab_obj = VoiceTab(voice_frame, self.settings, self.tts_engine, self.audio_router, self._on_change_placeholder, parent_window=self.window)
         self.tabs.append(self.voice_tab_obj)
 
         # Audio Output Tab
@@ -171,6 +171,11 @@ def _create_window(self):
         self.advanced_tab_obj = AdvancedTab(advanced_frame, self.settings, self.tts_engine, self.audio_router, self._on_change_placeholder)
         self.tabs.append(self.advanced_tab_obj)
 
+        # TTS Provider Tab
+        provider_frame = self.tabview.add("TTS Provider")
+        self.provider_tab_obj = TTSProviderTab(provider_frame, self.settings, self.tts_engine, self.audio_router, self._on_change_placeholder)
+        self.tabs.append(self.provider_tab_obj)
+
         # Buttons frame - fixed at bottom with standardized padding
         self.buttons_frame = ctk.CTkFrame(self.window, fg_color="transparent")
         self.buttons_frame.grid(row=1, column=0, padx=0, pady=(SPACING_SM, SPACING_SM), sticky="ew")

src/tts/providers/__init__.py

@@ -2,11 +2,11 @@
 TTS Provider Abstraction Layer
 
 This module defines the abstract base class for TTS providers and provides
-a common interface for TTS services like Edge TTS.
+a common interface for TTS services like Edge TTS and Piper TTS.
 """
 
 from abc import ABC, abstractmethod
-from typing import Dict, List, Any
+from typing import Dict, List, Any, Optional
 
 
 class TTSProvider(ABC):
@@ -54,4 +54,10 @@ def clear_cache(self) -> None:
         """Clear any cached data (voices, etc.)."""
         pass
 
+    def get_default_voice(self) -> Optional[str]:
+        """Return the default voice identifier for this provider.
 
+        Override in subclasses to provide a sensible provider-specific default.
+        Returns None if no default is defined.
+        """
+        return None

src/tts/providers/edge_tts_provider.py

@@ -294,3 +294,7 @@ def clear_cache(self) -> None:
         """Clear the voice cache."""
         self._voice_cache = None
         self._cache_time = 0
+
+    def get_default_voice(self) -> str:
+        """Return the default Edge TTS voice."""
+        return "en-US-AriaNeural"