whisper-key-local 0.4.0__tar.gz → 0.5.1__tar.gz

{whisper_key_local-0.4.0 → whisper_key_local-0.5.1}/PKG-INFO

@@ -1,11 +1,12 @@
 Metadata-Version: 2.4
 Name: whisper-key-local
-Version: 0.4.0
+Version: 0.5.1
 Summary: Local faster-whisper speech-to-text app with global hotkeys for Windows
 Author-email: Pin Wang <pinwang@gmail.com>
 Requires-Python: >=3.11
 Description-Content-Type: text/markdown
-Requires-Dist: faster-whisper>=1.1.1
+Requires-Dist: faster-whisper>=1.2.1
+Requires-Dist: ctranslate2>=4.6.3
 Requires-Dist: numpy>=1.24.0
 Requires-Dist: scipy>=1.11.0
 Requires-Dist: sounddevice>=0.4.6
@@ -22,6 +23,8 @@ Requires-Dist: hf-xet>=1.1.5
 
 Global hotkey to start/stop recording and auto-paste transcription wherever your cursor is.
 
+Questions or ideas? [Discord Server](https://discord.gg/uZnXV8snhz)
+
 ## 🎯 Features
 
 - **Global Hotkey**: Press `Ctrl+Win` to start recording, `Ctrl` to stop

{whisper_key_local-0.4.0 → whisper_key_local-0.5.1}/README.md

@@ -2,6 +2,8 @@
 
 Global hotkey to start/stop recording and auto-paste transcription wherever your cursor is.
 
+Questions or ideas? [Discord Server](https://discord.gg/uZnXV8snhz)
+
 ## 🎯 Features
 
 - **Global Hotkey**: Press `Ctrl+Win` to start recording, `Ctrl` to stop

{whisper_key_local-0.4.0 → whisper_key_local-0.5.1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "whisper-key-local"
-version = "0.4.0"
+version = "0.5.1"
 description = "Local faster-whisper speech-to-text app with global hotkeys for Windows"
 readme = "README.md"
 authors = [
@@ -12,7 +12,8 @@ authors = [
 ]
 requires-python = ">=3.11"
 dependencies = [
-    "faster-whisper>=1.1.1",
+    "faster-whisper>=1.2.1",
+    "ctranslate2>=4.6.3",
     "numpy>=1.24.0",
     "scipy>=1.11.0",
     "sounddevice>=0.4.6",

whisper_key_local-0.5.1/src/whisper_key/assets/version.txt

@@ -0,0 +1 @@
+0.5.1

{whisper_key_local-0.4.0 → whisper_key_local-0.5.1}/src/whisper_key/clipboard_manager.py

@@ -6,14 +6,18 @@ import pyperclip
 import win32gui
 import pyautogui
 
+from .utils import parse_hotkey
+
 pyautogui.FAILSAFE = True  # Enable "move mouse to corner to abort automation"
 
-class ClipboardManager:    
-    def __init__(self, key_simulation_delay, auto_paste, preserve_clipboard):
+class ClipboardManager:
+    def __init__(self, key_simulation_delay, auto_paste, preserve_clipboard, paste_hotkey):
         self.logger = logging.getLogger(__name__)
         self.key_simulation_delay = key_simulation_delay
         self.auto_paste = auto_paste
         self.preserve_clipboard = preserve_clipboard
+        self.paste_hotkey = paste_hotkey
+        self.paste_keys = parse_hotkey(paste_hotkey)
         self._configure_pyautogui_timing()
         self._test_clipboard_access()
         self._print_status()
@@ -31,11 +35,11 @@ class ClipboardManager:
             raise
     
     def _print_status(self):
+        hotkey_display = self.paste_hotkey.upper()
         if self.auto_paste:
-            method_name = "key simulation (CTRL+V)"
-            print(f"   ✓ Auto-paste is ENABLED using {method_name}")
+            print(f"   ✓ Auto-paste is ENABLED using key simulation ({hotkey_display})")
         else:
-            print("   ✗ Auto-paste is DISABLED - paste manually with Ctrl+V")
+            print(f"   ✗ Auto-paste is DISABLED - paste manually with {hotkey_display}")
     
     def copy_text(self, text: str) -> bool:
         if not text:
@@ -106,7 +110,7 @@ class ClipboardManager:
             if not self.copy_text(text):
                 return False
                       
-            pyautogui.hotkey('ctrl', 'v')
+            pyautogui.hotkey(*self.paste_keys)
 
             print(f"   ✓ Auto-pasted via key simulation")
 

{whisper_key_local-0.4.0 → whisper_key_local-0.5.1}/src/whisper_key/config.defaults.yaml

@@ -7,24 +7,11 @@
 # Personal settings are stored at:
 # %APPDATA%\Roaming\whisperkey\user_settings.yaml
 
-# =============================================================================
-# WHISPER AI MODEL SETTINGS
-# =============================================================================
-whisper:
-  # Model size - affects accuracy vs speed tradeoff
-  # Options: "tiny", "base", "small", "medium", "large", "tiny.en", "base.en", "small.en", "medium.en"
-  # Multilingual models:
-  # - tiny: ~39MB, fastest, good for most use cases
-  # - base: ~74MB, better accuracy, still fast
-  # - small: ~244MB, high accuracy, slower
-  # - medium: ~769MB, very high accuracy, much slower
-  # - large: ~1550MB, best accuracy, very slow (not recommended for real-time)
-  # English-only models (more accurate, slightly faster):
-  # - tiny.en: 39MB, 
-  # - base.en: 74MB, 
-  # - small.en: 244MB, 
-  # - medium.en: 769MB
-  model_size: base
+whisper: # Whisper AI Model Settings
+
+  # Model selection
+  # See "models" below for more information
+  model: tiny
   
   # Processing device - where the AI runs
   # Options: "cpu", "cuda" (for NVIDIA GPUs)
@@ -46,10 +33,60 @@ whisper:
   # Transcription quality settings
   beam_size: 5  # Higher = more accurate but slower (1-10)
 
-# =============================================================================
-# HOTKEY CONFIGURATION
-# =============================================================================
-hotkey:
+  # Available models (set enabled: false to hide from menu)
+  # To add custom models (CTranslate2 format), add entry with source (HuggingFace or local path)
+  # Example: my-model: {source: "username/repo-ct2", label: "My Model", group: custom}
+  models:
+    # Official whisper models
+    tiny:
+      label: "Tiny (76MB, fastest)"
+      group: official
+      enabled: true
+    base:
+      label: "Base (145MB, balanced)"
+      group: official
+      enabled: true
+    small:
+      label: "Small (484MB, accurate)"
+      group: official
+      enabled: true
+    medium:
+      label: "Medium (1.5GB, very accurate)"
+      group: official
+      enabled: true
+    large: # Uses the large-v3 version
+      label: "Large (3.1GB, best accuracy)"
+      group: official
+      enabled: true
+    large-v3-turbo: # 8x faster and only slightly less accurate (transcription optimized)
+      label: "Large-V3-Turbo (1.6GB, newest)"
+      group: official
+      enabled: true
+
+    # Models with native faster-whisper support
+    tiny.en:
+      label: "Tiny.En (English)"
+      group: custom
+      enabled: true
+    base.en:
+      label: "Base.En (English)"
+      group: custom
+      enabled: true
+    small.en:
+      label: "Small.En (English)"
+      group: custom
+      enabled: true
+    medium.en:
+      label: "Medium.En (English)"
+      group: custom
+      enabled: true
+    distil-large-v3.5: # About 1.5x faster than large-v3-turbo
+      source: distil-whisper/distil-large-v3.5-ct2
+      label: "Distil-Large-V3.5 (English)"
+      group: custom
+
+hotkey: # Hotkey Configuration
+
   # Key combination to start/stop recording
   # Format: modifier+modifier+key (use lowercase)
   # Common modifiers: ctrl, shift, alt, win
@@ -79,10 +116,8 @@ hotkey:
   # Examples: "esc", "ctrl+c", "shift+esc" 
   cancel_combination: esc
 
-# =============================================================================
-# VOICE ACTIVITY DETECTION (VAD) SETTINGS
-# =============================================================================
-vad:
+vad: # Voice Activity Detection (VAD)
+
   # Voice Activity Detection (VAD) Pre-check
   # Uses TEN VAD to detect speech before transcription (prevents hallucinations on silence)
   # NOTE: This is separate from faster-whisper's built-in VAD
@@ -99,10 +134,8 @@ vad:
   vad_realtime_enabled: true  # Enables continuous silence detection and automatic recording stop
   vad_silence_timeout_seconds: 30.0  # Auto-stop recording after this many seconds of silence
 
-# =============================================================================
-# AUDIO RECORDING SETTINGS
-# =============================================================================
-audio:
+audio: # Audio Recording Settings
+
   # Sample rate is fixed at 16000 Hz for optimal Whisper and TEN VAD performance
 
   # Audio host API selection
@@ -132,28 +165,32 @@ audio:
   # are plugged/unplugged. Use system tray to select device - it will save the ID here.
   input_device: "default"
 
-# =============================================================================
-# CLIPBOARD BEHAVIOR
-# =============================================================================
-clipboard:
+clipboard: # Clipboard Behavior
+
   # Automatically paste after transcription
   # true = paste immediately to active window
   # false = only copy to clipboard (paste manually with Ctrl+V)
   auto_paste: true
-  
+
+  # Key combination to simulate paste
+  # Format: modifier+key (use lowercase)
+  # Examples:
+  #   - "ctrl+v"         (standard paste)
+  #   - "ctrl+shift+v"   (plain text paste in some apps)
+  #   - "shift+insert"   (terminal paste)
+  paste_hotkey: ctrl+v
+
   # Preserve existing clipboard content when pasting
   # true = save current clipboard, paste transcription, then restore original clipboard
   # false = replace clipboard with transcription (default behavior)
   preserve_clipboard: true
-  
+
   # Delay for key simulation operations (seconds)
   # Small delay ensures operations complete reliably
   key_simulation_delay: 0.05
 
-# =============================================================================
-# LOGGING CONFIGURATION
-# =============================================================================
-logging:
+logging: # Logging Configuration
+
   # Log level - how much detail to log
   # Options: "DEBUG", "INFO", "WARNING", "ERROR", "CRITICAL"
   # - DEBUG: Very detailed (for troubleshooting)
@@ -173,10 +210,8 @@ logging:
     level: WARNING  # Console log level (DEBUG, INFO, WARNING, ERROR, CRITICAL) - WARNING shows important messages only
 
 
-# =============================================================================
-# AUDIO FEEDBACK SETTINGS
-# =============================================================================
-audio_feedback:
+audio_feedback: # Audio Feedback Settings
+
   # Enable/disable audio feedback sounds
   # true = play sounds when recording starts/stops
   # false = silent operation
@@ -188,10 +223,8 @@ audio_feedback:
   stop_sound: assets/sounds/record_stop.wav
   cancel_sound: assets/sounds/record_cancel.wav
 
-# =============================================================================
-# SYSTEM TRAY SETTINGS
-# =============================================================================
-system_tray:
+system_tray: # System Tray Settings
+
   # Enable/disable system tray icon
   # true = show icon in system tray with status and menu
   # false = run without system tray (console mode only)
@@ -200,10 +233,8 @@ system_tray:
   # Tooltip text when hovering over tray icon
   tooltip: Whisper Key
 
-# =============================================================================
-# CONSOLE VISIBILITY SETTINGS
-# =============================================================================
-console:
+console: # Console Visibility Settings
+
   # Start with console hidden to system tray
   # Note: This setting is ignored in CLI mode
   start_hidden: false

{whisper_key_local-0.4.0 → whisper_key_local-0.5.1}/src/whisper_key/config_manager.py

@@ -210,6 +210,10 @@ class ConfigManager:
     def get_console_config(self) -> Dict[str, Any]:
         return self.config.get('console', {}).copy()
 
+    def get_log_file_path(self) -> str:
+        log_filename = self.config['logging']['file']['filename']
+        return os.path.join(get_user_app_data_path(), log_filename)
+
     def get_setting(self, section: str, key: str) -> Any:
         return self.config[section][key]
     
@@ -217,25 +221,30 @@ class ConfigManager:
         yaml = YAML()
         yaml.preserve_quotes = True
         yaml.indent(mapping=2, sequence=4, offset=2)
-        
+
         temp_output = StringIO()
         yaml.dump(config_data, temp_output)
         lines = temp_output.getvalue().split('\n')
-        
-        # Find end of header - first blank line is the cutoff
-        data_start = 0
+
+        content_start = 0
         for i, line in enumerate(lines):
-            if not line.strip():  # Empty line found
-                data_start = i
+            stripped = line.strip()
+            if stripped and not stripped.startswith('#'):
+                content_start = i
                 break
-        
-        user_config = []
-        user_config.append("# =============================================================================")
-        user_config.append("# WHISPER KEY - PERSONAL CONFIGURATION")
-        user_config.append("# =============================================================================")
-        user_config.extend(lines[data_start:])
-        
-        return '\n'.join(user_config)
+
+        content_lines = lines[content_start:]
+
+        header = [
+            "# =============================================================================",
+            "# WHISPER KEY - PERSONAL CONFIGURATION",
+            "# =============================================================================",
+            "# Edit this file to customize your settings",
+            "# Save and restart Whisper Key for changes to take effect",
+            ""
+        ]
+
+        return '\n'.join(header + content_lines)
 
     def save_config_to_user_settings_file(self):
         try:
@@ -339,8 +348,6 @@ class ConfigValidator:
         self.config = config
         self.default_config = default_config
         
-        self._validate_enum('whisper.model_size', 
-                            ['tiny', 'base', 'small', 'medium', 'large', 'tiny.en', 'base.en', 'small.en', 'medium.en'])
         self._validate_enum('whisper.device', ['cpu', 'cuda'])
         self._validate_enum('whisper.compute_type', ['int8', 'float16', 'float32'])
         
@@ -354,6 +361,7 @@ class ConfigValidator:
         self._validate_boolean('clipboard.auto_paste')
         self._validate_boolean('clipboard.preserve_clipboard')
         self._validate_numeric_range('clipboard.key_simulation_delay', min_val=0, description='key simulation delay')
+        self._validate_hotkey_string('clipboard.paste_hotkey')
         
         self._validate_boolean('hotkey.stop_with_modifier_enabled')
         self._validate_boolean('hotkey.auto_enter_enabled')

{whisper_key_local-0.4.0 → whisper_key_local-0.5.1}/src/whisper_key/main.py

@@ -20,6 +20,7 @@ from .system_tray import SystemTray
 from .audio_feedback import AudioFeedback
 from .console_manager import ConsoleManager
 from .instance_manager import guard_against_multiple_instances
+from .model_registry import ModelRegistry
 from .utils import beautify_hotkey, get_user_app_data_path, get_version
 
 def is_built_executable():
@@ -82,21 +83,23 @@ def setup_vad(vad_config):
         vad_silence_timeout_seconds=vad_config['vad_silence_timeout_seconds']
     )
 
-def setup_whisper_engine(whisper_config, vad_manager):
+def setup_whisper_engine(whisper_config, vad_manager, model_registry):
     return WhisperEngine(
-        model_size=whisper_config['model_size'],
+        model_key=whisper_config['model'],
         device=whisper_config['device'],
         compute_type=whisper_config['compute_type'],
         language=whisper_config['language'],
         beam_size=whisper_config['beam_size'],
-        vad_manager=vad_manager
+        vad_manager=vad_manager,
+        model_registry=model_registry
     )
 
 def setup_clipboard_manager(clipboard_config):
     return ClipboardManager(
         key_simulation_delay=clipboard_config['key_simulation_delay'],
         auto_paste=clipboard_config['auto_paste'],
-        preserve_clipboard=clipboard_config['preserve_clipboard']
+        preserve_clipboard=clipboard_config['preserve_clipboard'],
+        paste_hotkey=clipboard_config['paste_hotkey']
     )
 
 def setup_audio_feedback(audio_feedback_config):
@@ -113,11 +116,12 @@ def setup_console_manager(console_config, is_executable_mode):
         is_executable_mode=is_executable_mode
     )
 
-def setup_system_tray(tray_config, config_manager, state_manager=None):
+def setup_system_tray(tray_config, config_manager, state_manager, model_registry):
     return SystemTray(
         state_manager=state_manager,
         tray_config=tray_config,
-        config_manager=config_manager
+        config_manager=config_manager,
+        model_registry=model_registry
     )
 
 def setup_signal_handlers(shutdown_event):
@@ -179,8 +183,9 @@ def main():
         is_executable = is_built_executable()
         console_manager = setup_console_manager(console_config, is_executable)
 
+        model_registry = ModelRegistry(whisper_config.get('models', {}))
         vad_manager = setup_vad(vad_config)
-        whisper_engine = setup_whisper_engine(whisper_config, vad_manager)
+        whisper_engine = setup_whisper_engine(whisper_config, vad_manager, model_registry)
         clipboard_manager = setup_clipboard_manager(clipboard_config)
         audio_feedback = setup_audio_feedback(audio_feedback_config)
 
@@ -195,7 +200,7 @@ def main():
             vad_manager=vad_manager
         )
         audio_recorder = setup_audio_recorder(audio_config, state_manager, vad_manager)
-        system_tray = setup_system_tray(tray_config, config_manager, state_manager)
+        system_tray = setup_system_tray(tray_config, config_manager, state_manager, model_registry)
         state_manager.attach_components(audio_recorder, system_tray)
         
         hotkey_listener = setup_hotkey_listener(hotkey_config, state_manager)

whisper_key_local-0.5.1/src/whisper_key/model_registry.py

@@ -0,0 +1,77 @@
+import os
+from faster_whisper.utils import _MODELS
+
+
+class ModelRegistry:
+    DEFAULT_CACHE_PREFIX = "models--Systran--faster-whisper-"
+
+    def __init__(self, models_config: dict):
+        self.models = {}
+        for key, config in models_config.items():
+            if isinstance(config, dict):
+                self.models[key] = ModelDefinition(key, config)
+
+    def get_model(self, key: str):
+        return self.models.get(key)
+
+    def get_source(self, key: str) -> str:
+        model = self.get_model(key)
+        return model.source if model else key
+
+    def get_cache_folder(self, key: str) -> str:
+        model = self.get_model(key)
+        if not model:
+            return f"{self.DEFAULT_CACHE_PREFIX}{key}"
+        return model.cache_folder
+
+    def get_models_by_group(self, group: str) -> list:
+        return [m for m in self.models.values() if m.group == group and m.enabled]
+
+    def get_groups_ordered(self) -> list:
+        return ["official", "custom"]
+
+    def get_hf_cache_path(self) -> str:
+        userprofile = os.environ.get('USERPROFILE')
+        if userprofile:
+            return os.path.join(userprofile, '.cache', 'huggingface', 'hub')
+        return os.path.join(os.path.expanduser('~'), '.cache', 'huggingface', 'hub')
+
+    def is_model_cached(self, key: str) -> bool:
+        model = self.get_model(key)
+        if model and model.is_local_path:
+            return os.path.exists(os.path.join(model.source, 'model.bin'))
+        cache_folder = self.get_cache_folder(key)
+        if not cache_folder:
+            return False
+        return os.path.exists(os.path.join(self.get_hf_cache_path(), cache_folder))
+
+
+class ModelDefinition:
+    def __init__(self, key: str, config: dict):
+        self.key = key
+        self.source = config.get("source", key)
+        self.label = config.get("label", key.title())
+        self.group = config.get("group", "custom")
+        self.enabled = config.get("enabled", True)
+        self.is_local_path = self._check_is_local_path()
+        self.cache_folder = self._derive_cache_folder()
+
+    def _check_is_local_path(self) -> bool:
+        if self.source.startswith("\\\\") or (len(self.source) > 2 and self.source[1] == ":"):
+            return True
+        if "/" in self.source:
+            return os.path.exists(self.source)
+        return False
+
+    def _derive_cache_folder(self) -> str:
+        if self.is_local_path:
+            return None
+
+        if "/" in self.source:
+            return "models--" + self.source.replace("/", "--")
+
+        if self.source in _MODELS:
+            repo = _MODELS[self.source]
+            return "models--" + repo.replace("/", "--")
+
+        return f"{ModelRegistry.DEFAULT_CACHE_PREFIX}{self.source}"

{whisper_key_local-0.4.0 → whisper_key_local-0.5.1}/src/whisper_key/state_manager.py

@@ -160,7 +160,7 @@ class StateManager:
 
             if pending_model:
                 self.logger.info(f"Executing pending model change to: {pending_model}")
-                print(f"🔄 Processing complete, now switching to {pending_model} model...")
+                print(f"🔄 Processing complete, now switching to [{pending_model}] model...")
                 self._execute_model_change(pending_model)
                 self._pending_model_change = None
 
@@ -226,10 +226,10 @@ class StateManager:
             else:
                 return "idle"
     
-    def request_model_change(self, new_model_size: str) -> bool:
+    def request_model_change(self, new_model_key: str) -> bool:
         current_state = self.get_current_state()
         
-        if new_model_size == self.whisper_engine.model_size:
+        if new_model_key == self.whisper_engine.model_key:
             return True
         
         if current_state == "model_loading":
@@ -237,18 +237,18 @@ class StateManager:
             return False
         
         if current_state == "recording":
-            print(f"🎤 Cancelling recording to switch to {new_model_size} model...")
+            print(f"🎤 Cancelling recording to switch to [{new_model_key}] model...")
             self.cancel_active_recording()
-            self._execute_model_change(new_model_size)
+            self._execute_model_change(new_model_key)
             return True
         
         if current_state == "processing":
-            print(f"⏳ Queueing model change to {new_model_size} until transcription completes...")
-            self._pending_model_change = new_model_size
+            print(f"⏳ Queueing model change to [{new_model_key}] until transcription completes...")
+            self._pending_model_change = new_model_key
             return True
         
         if current_state == "idle":
-            self._execute_model_change(new_model_size)
+            self._execute_model_change(new_model_key)
             return True
         
         self.logger.warning(f"Unexpected state for model change: {current_state}")
@@ -261,10 +261,10 @@ class StateManager:
     def show_console(self):
         self.console_manager.show_console()
 
-    def _execute_model_change(self, new_model_size: str):
+    def _execute_model_change(self, new_model_key: str):
         def progress_callback(message: str):
             if "ready" in message.lower() or "already loaded" in message.lower():
-                print(f"✅ Successfully switched to {new_model_size} model")
+                print(f"✅ Successfully switched to [{new_model_key}] model")
                 self.set_model_loading(False)
             elif "failed" in message.lower():
                 print(f"❌ Failed to change model: {message}")
@@ -275,9 +275,9 @@ class StateManager:
         
         try:
             self.set_model_loading(True)
-            print(f"🔄 Switching to {new_model_size} model...")
+            print(f"🔄 Switching to [{new_model_key}] model...")
             
-            self.whisper_engine.change_model(new_model_size, progress_callback)
+            self.whisper_engine.change_model(new_model_key, progress_callback)
             
         except Exception as e:
             self.logger.error(f"Failed to initiate model change: {e}")

{whisper_key_local-0.4.0 → whisper_key_local-0.5.1}/src/whisper_key/system_tray.py

@@ -20,15 +20,17 @@ if TYPE_CHECKING:
     from .state_manager import StateManager
     from .config_manager import ConfigManager
 
-class SystemTray:   
+class SystemTray:
     def __init__(self,
                  state_manager: 'StateManager',
                  tray_config: dict = None,
-                 config_manager: Optional['ConfigManager'] = None):
+                 config_manager: Optional['ConfigManager'] = None,
+                 model_registry = None):
 
         self.state_manager = state_manager
         self.tray_config = tray_config or {}
         self.config_manager = config_manager
+        self.model_registry = model_registry
         self.logger = logging.getLogger(__name__)
                
         self.icon = None  # pystray object, holds menu, state, etc.
@@ -91,13 +93,49 @@ class SystemTray:
 
         return icon
     
+    def _build_model_menu_items(self, current_model: str, is_model_loading: bool) -> list:
+        items = []
+
+        if not self.model_registry:
+            return items
+
+        def make_model_selector(model_key):
+            return lambda icon, item: self._select_model(model_key)
+
+        def make_is_current(model_key):
+            return lambda item: model_key == current_model
+
+        def model_selection_enabled(item):
+            return not is_model_loading
+
+        first_group = True
+        for group in self.model_registry.get_groups_ordered():
+            models = self.model_registry.get_models_by_group(group)
+            if not models:
+                continue
+
+            if not first_group:
+                items.append(pystray.Menu.SEPARATOR)
+            first_group = False
+
+            for model in models:
+                items.append(pystray.MenuItem(
+                    model.label,
+                    make_model_selector(model.key),
+                    radio=True,
+                    checked=make_is_current(model.key),
+                    enabled=model_selection_enabled
+                ))
+
+        return items
+
     def _create_menu(self):
         try:
             app_state = self.state_manager.get_application_state()
             is_model_loading = app_state.get('model_loading', False)
 
             auto_paste_enabled = self.config_manager.get_setting('clipboard', 'auto_paste')
-            current_model = self.config_manager.get_setting('whisper', 'model_size')
+            current_model = self.config_manager.get_setting('whisper', 'model')
 
             available_hosts = self.state_manager.get_available_audio_hosts()
             current_host = self.state_manager.get_current_audio_host()
@@ -146,26 +184,12 @@ class SystemTray:
                         )
                     )
 
-            def is_current_model(model_name):
-                return model_name == current_model
-
-            def model_selection_enabled():
-                return not is_model_loading
-
-            model_sub_menu_items = [
-                pystray.MenuItem("Tiny (75MB, fastest)", lambda icon, item: self._select_model("tiny"), radio=True, checked=lambda item: is_current_model("tiny"), enabled=model_selection_enabled()),
-                pystray.MenuItem("Base (142MB, balanced)", lambda icon, item: self._select_model("base"), radio=True, checked=lambda item: is_current_model("base"), enabled=model_selection_enabled()),
-                pystray.MenuItem("Small (466MB, accurate)", lambda icon, item: self._select_model("small"), radio=True, checked=lambda item: is_current_model("small"), enabled=model_selection_enabled()),
-                pystray.MenuItem("Medium (1.5GB, very accurate)", lambda icon, item: self._select_model("medium"), radio=True, checked=lambda item: is_current_model("medium"), enabled=model_selection_enabled()),
-                pystray.MenuItem("Large (2.9GB, best accuracy)", lambda icon, item: self._select_model("large"), radio=True, checked=lambda item: is_current_model("large"), enabled=model_selection_enabled()),
-                pystray.Menu.SEPARATOR,
-                pystray.MenuItem("Tiny.En (English only)", lambda icon, item: self._select_model("tiny.en"), radio=True, checked=lambda item: is_current_model("tiny.en"), enabled=model_selection_enabled()),
-                pystray.MenuItem("Base.En (English only)", lambda icon, item: self._select_model("base.en"), radio=True, checked=lambda item: is_current_model("base.en"), enabled=model_selection_enabled()),
-                pystray.MenuItem("Small.En (English only)", lambda icon, item: self._select_model("small.en"), radio=True, checked=lambda item: is_current_model("small.en"), enabled=model_selection_enabled()),
-                pystray.MenuItem("Medium.En (English only)", lambda icon, item: self._select_model("medium.en"), radio=True, checked=lambda item: is_current_model("medium.en"), enabled=model_selection_enabled())
-            ]
+            model_sub_menu_items = self._build_model_menu_items(current_model, is_model_loading)
 
             menu_items = [
+                pystray.MenuItem("View Log", self._view_log_file),
+                pystray.MenuItem("Advanced Settings", self._open_config_file),
+                pystray.Menu.SEPARATOR,
                 pystray.MenuItem(
                     "Audio Host",
                     pystray.Menu(*audio_host_items)
@@ -207,22 +231,38 @@ class SystemTray:
     def _show_console(self, icon=None, item=None):
         self.state_manager.show_console()
 
+    def _view_log_file(self, icon=None, item=None):
+        try:
+            print("⚙️ Opening log file...")
+            log_path = self.config_manager.get_log_file_path()
+            os.startfile(log_path)
+        except Exception as e:
+            self.logger.error(f"Failed to open log file: {e}")
+
+    def _open_config_file(self, icon=None, item=None):
+        try:
+            print("⚙️ Opening settings...")
+            config_path = self.config_manager.user_settings_path
+            os.startfile(config_path)
+        except Exception as e:
+            self.logger.error(f"Failed to open config file: {e}")
+
     def _set_transcription_mode(self, auto_paste: bool):        
         self.state_manager.update_transcription_mode(auto_paste)
         self.icon.menu = self._create_menu()
 
-    def _select_model(self, model_size: str):
+    def _select_model(self, model_key: str):
         try:
-            success = self.state_manager.request_model_change(model_size)
+            success = self.state_manager.request_model_change(model_key)
 
             if success:
-                self.config_manager.update_user_setting('whisper', 'model_size', model_size)
+                self.config_manager.update_user_setting('whisper', 'model', model_key)
                 self.icon.menu = self._create_menu()
             else:
-                self.logger.warning(f"Request to change model to {model_size} was not accepted")
+                self.logger.warning(f"Request to change model to {model_key} was not accepted")
 
         except Exception as e:
-            self.logger.error(f"Error selecting model {model_size}: {e}")
+            self.logger.error(f"Error selecting model {model_key}: {e}")
 
     def _select_audio_host(self, host_name: str):
         try:

{whisper_key_local-0.4.0 → whisper_key_local-0.5.1}/src/whisper_key/utils.py

@@ -20,9 +20,14 @@ class OptionalComponent:
 def beautify_hotkey(hotkey_string: str) -> str:
     if not hotkey_string:
         return ""
-    
+
     return hotkey_string.replace('+', '+').upper()
 
+def parse_hotkey(hotkey_string: str) -> list:
+    if not hotkey_string:
+        return []
+    return hotkey_string.lower().split('+')
+
 def is_installed_package():
     # Check if running from an installed package
     return 'site-packages' in __file__

{whisper_key_local-0.4.0 → whisper_key_local-0.5.1}/src/whisper_key/whisper_engine.py

@@ -1,111 +1,109 @@
 import logging
-import os
 import time
 import threading
 from typing import Optional, Callable
 
 import numpy as np
 from faster_whisper import WhisperModel
-from .utils import OptionalComponent
 
-class WhisperEngine:
-    MODEL_CACHE_PREFIX = "models--Systran--faster-whisper-"  # file prefix for hugging-face model
 
+class WhisperEngine:
     def __init__(self,
-                 model_size: str = "tiny",
+                 model_key: str = "tiny",
                  device: str = "cpu",
                  compute_type: str = "int8",
                  language: str = None,
                  beam_size: int = 5,
-                 vad_manager = None):
-        
-        self.model_size = model_size
+                 vad_manager = None,
+                 model_registry = None):
+
+        self.model_key = model_key
         self.device = device
         self.compute_type = compute_type
         self.language = None if language == 'auto' else language
         self.beam_size = beam_size
         self.model = None
         self.logger = logging.getLogger(__name__)
+        self.registry = model_registry
 
         self._loading_thread = None
         self._progress_callback = None
 
         self.vad_manager = vad_manager
-        
+
         self._load_model()
     
-    def _get_cache_directory(self):
-        userprofile = os.getenv('USERPROFILE')
-        if not userprofile:
-            home = os.path.expanduser('~')
-            userprofile = home
-        
-        cache_dir = os.path.join(userprofile, '.cache', 'huggingface', 'hub')
-        return cache_dir
-    
-    def _is_model_cached(self, model_size=None):
-        if model_size is None:
-            model_size = self.model_size
-        cache_dir = self._get_cache_directory()
-        model_folder = f"{self.MODEL_CACHE_PREFIX}{model_size}"
-        return os.path.exists(os.path.join(cache_dir, model_folder))
-    
+    def _get_model_source(self, model_key: str) -> str:
+        if self.registry:
+            return self.registry.get_source(model_key)
+        return model_key
+
+    def _is_model_cached(self, model_key: str = None) -> bool:
+        if model_key is None:
+            model_key = self.model_key
+        if self.registry:
+            return self.registry.is_model_cached(model_key)
+        return False
+
     def _load_model(self):
         try:
-            print(f"🧠 Loading Whisper AI model [{self.model_size}]...")
-            
+            print(f"🧠 Loading Whisper AI model [{self.model_key}]...")
+
             was_cached = self._is_model_cached()
             if not was_cached:
                 print("Downloading model, this may take a few minutes....")
-            
+
+            model_source = self._get_model_source(self.model_key)
             self.model = WhisperModel(
-                self.model_size,
+                model_source,
                 device=self.device,
                 compute_type=self.compute_type
             )
-            
+
             if not was_cached:
-                print("\n") # Workaround for download status bar misplacement
+                print("\n")  # Workaround for download status bar misplacement
+
+            print(f"   ✓ Whisper model [{self.model_key}] ready!")
 
-            print(f"   ✓ Whisper model [{self.model_size}] ready!")
-            
         except Exception as e:
             self.logger.error(f"Failed to load Whisper model: {e}")
             raise
     
     def _load_model_async(self,
-                          new_model_size: str,
+                          new_model_key: str,
                           progress_callback: Optional[Callable[[str], None]] = None):
         def _background_loader():
-            try:             
+            try:
                 if progress_callback:
                     progress_callback("Checking model cache...")
-                
-                old_model_size = self.model_size
-                was_cached = self._is_model_cached(new_model_size)
-                
+
+                old_model_key = self.model_key
+                was_cached = self._is_model_cached(new_model_key)
+
                 if progress_callback:
                     if was_cached:
                         progress_callback("Loading cached model...")
                     else:
                         progress_callback("Downloading model...")
-                               
-                self.logger.info(f"Loading Whisper model: {new_model_size} (async)")
 
+                self.logger.info(f"Loading Whisper model: {new_model_key} (async)")
+
+                model_source = self._get_model_source(new_model_key)
                 new_model = WhisperModel(
-                    new_model_size,
+                    model_source,
                     device=self.device,
                     compute_type=self.compute_type
                 )
-                
+
                 self.model = new_model
-                self.logger.info(f"Whisper model [{new_model_size}] loaded successfully (async)")
-                
+                self.model_key = new_model_key
+                self.logger.info(f"Whisper model [{new_model_key}] loaded successfully (async)")
+
                 if progress_callback:
                     progress_callback("Model ready!")
-                
+
             except Exception as e:
-                self.model_size = old_model_size
+                self.model_key = old_model_key
                 self.logger.error(f"Failed to load Whisper model async: {e}")
                 if progress_callback:
                     progress_callback(f"Failed to load model: {e}")
@@ -113,7 +111,7 @@ class WhisperEngine:
             finally:
                 self._loading_thread = None
                 self._progress_callback = None
-        
+
         if self._loading_thread and self._loading_thread.is_alive():
             self.logger.warning("Model loading already in progress, ignoring new request")
             return
@@ -188,13 +186,13 @@ class WhisperEngine:
     
     
     def change_model(self,
-                     new_model_size: str,
+                     new_model_key: str,
                      progress_callback: Optional[Callable[[str], None]] = None):
         
-        if new_model_size == self.model_size:
+        if new_model_key == self.model_key:
             if progress_callback:
                 progress_callback("Model already loaded")
             return
         
-        self._load_model_async(new_model_size, progress_callback)
+        self._load_model_async(new_model_key, progress_callback)
     

{whisper_key_local-0.4.0 → whisper_key_local-0.5.1}/src/whisper_key_local.egg-info/PKG-INFO

@@ -1,11 +1,12 @@
 Metadata-Version: 2.4
 Name: whisper-key-local
-Version: 0.4.0
+Version: 0.5.1
 Summary: Local faster-whisper speech-to-text app with global hotkeys for Windows
 Author-email: Pin Wang <pinwang@gmail.com>
 Requires-Python: >=3.11
 Description-Content-Type: text/markdown
-Requires-Dist: faster-whisper>=1.1.1
+Requires-Dist: faster-whisper>=1.2.1
+Requires-Dist: ctranslate2>=4.6.3
 Requires-Dist: numpy>=1.24.0
 Requires-Dist: scipy>=1.11.0
 Requires-Dist: sounddevice>=0.4.6
@@ -22,6 +23,8 @@ Requires-Dist: hf-xet>=1.1.5
 
 Global hotkey to start/stop recording and auto-paste transcription wherever your cursor is.
 
+Questions or ideas? [Discord Server](https://discord.gg/uZnXV8snhz)
+
 ## 🎯 Features
 
 - **Global Hotkey**: Press `Ctrl+Win` to start recording, `Ctrl` to stop

{whisper_key_local-0.4.0 → whisper_key_local-0.5.1}/src/whisper_key_local.egg-info/SOURCES.txt

@@ -10,6 +10,7 @@ src/whisper_key/console_manager.py
 src/whisper_key/hotkey_listener.py
 src/whisper_key/instance_manager.py
 src/whisper_key/main.py
+src/whisper_key/model_registry.py
 src/whisper_key/state_manager.py
 src/whisper_key/system_tray.py
 src/whisper_key/utils.py

{whisper_key_local-0.4.0 → whisper_key_local-0.5.1}/src/whisper_key_local.egg-info/requires.txt

@@ -1,4 +1,5 @@
-faster-whisper>=1.1.1
+faster-whisper>=1.2.1
+ctranslate2>=4.6.3
 numpy>=1.24.0
 scipy>=1.11.0
 sounddevice>=0.4.6

whisper_key_local-0.4.0/src/whisper_key/assets/version.txt

@@ -1 +0,0 @@
-0.4.0