feat(setup): add ESC cancel, multi-language UI, and dynamic language …

KoeApp/Koe/AppDelegate/SPAppDelegate.m

@@ -294,6 +294,7 @@ - (void)hotkeyMonitorDidDetectHoldStart {
     NSLog(@"[Koe] Hold start detected");
     [self stopNumberKeyMonitoring];
     [self stopAnyKeyDismissMonitoring];
+    [self startEscapeKeyMonitoring];
     [self.overlayPanel hideTemplateButtons];
     self.showingError = NO;
     [self cancelPendingSessionEnd];
@@ -315,6 +316,7 @@ - (void)hotkeyMonitorDidDetectHoldStart {
 
 - (void)hotkeyMonitorDidDetectHoldEnd {
     NSLog(@"[Koe] Hold end detected");
+    [self stopEscapeKeyMonitoring];
     [self.cuePlayer playStop];
 
     // Keep recording for 300ms after Fn release to capture trailing speech,
@@ -335,6 +337,7 @@ - (void)hotkeyMonitorDidDetectTapStart {
     NSLog(@"[Koe] Tap start detected");
     [self stopNumberKeyMonitoring];
     [self stopAnyKeyDismissMonitoring];
+    [self startEscapeKeyMonitoring];
     [self.overlayPanel hideTemplateButtons];
     self.showingError = NO;
     [self cancelPendingSessionEnd];
@@ -355,6 +358,7 @@ - (void)hotkeyMonitorDidDetectTapStart {
 
 - (void)hotkeyMonitorDidDetectTapEnd {
     NSLog(@"[Koe] Tap end detected");
+    [self stopEscapeKeyMonitoring];
     [self.cuePlayer playStop];
 
     // Keep recording for 300ms after tap-end to capture trailing speech,
@@ -628,7 +632,9 @@ - (void)statusBarDidSelectSetupWizard {
         self.setupWizard.delegate = self;
         self.setupWizard.rustBridge = self.rustBridge;
     }
+    [NSApp activateIgnoringOtherApps:YES];
     [self.setupWizard showWindow:nil];
+    [self.setupWizard.window makeKeyAndOrderFront:nil];
 }
 
 - (void)statusBarDidSelectCheckForUpdates {
@@ -720,4 +726,33 @@ - (void)stopAnyKeyDismissMonitoring {
     self.hotkeyMonitor.anyKeyDismissHandler = nil;
 }
 
+#pragma mark - ESC Cancel Monitoring
+
+- (void)startEscapeKeyMonitoring {
+    // Default enabled: only disable if explicitly set to "false" in config.
+    char *raw = sp_config_get("hotkey.escape_cancel_enabled");
+    BOOL disabled = (raw && strcmp(raw, "false") == 0);
+    if (raw) sp_core_free_string(raw);
+    if (disabled) {
+        return;
+    }
+    __weak typeof(self) weakSelf = self;
+    self.hotkeyMonitor.escapeKeyHandler = ^{
+        __strong typeof(weakSelf) strongSelf = weakSelf;
+        if (!strongSelf) return;
+        NSLog(@"[Koe] ESC pressed — cancelling session");
+        [strongSelf stopEscapeKeyMonitoring];
+        [strongSelf cancelPendingSessionEnd];
+        [strongSelf.audioCaptureManager stopCapture];
+        [strongSelf.rustBridge cancelSession];
+        [strongSelf.hotkeyMonitor resetToIdle];
+        [strongSelf.overlayPanel updateState:@"idle"];
+        [strongSelf.statusBarManager updateState:@"idle"];
+    };
+}
+
+- (void)stopEscapeKeyMonitoring {
+    self.hotkeyMonitor.escapeKeyHandler = nil;
+}
+
 @end

KoeApp/Koe/Hotkey/SPHotkeyMonitor.h

@@ -56,4 +56,9 @@ typedef NS_ENUM(uint8_t, SPHotkeyMatchKind) {
 /// Used to dismiss the overlay when the user resumes typing after text insertion.
 @property (nonatomic, copy) void (^anyKeyDismissHandler)(void);
 
+/// Optional block called when ESC (keyCode 53) is pressed during an active recording session.
+/// The key event is NOT consumed — it passes through to the target app.
+/// Use this to cancel the current session without producing output.
+@property (nonatomic, copy) void (^escapeKeyHandler)(void);
+
 @end

KoeApp/Koe/Hotkey/SPHotkeyMonitor.m

@@ -104,6 +104,15 @@ static CGEventRef hotkeyEventCallback(CGEventTapProxy proxy,
             return monitor.canConsumeGlobalKeyEvents ? NULL : event;
         }
 
+        // ESC (keyCode 53) during recording → cancel session.
+        if (type == kCGEventKeyDown && keyCode == 53 && monitor.escapeKeyHandler) {
+            void (^handler)(void) = monitor.escapeKeyHandler;
+            dispatch_async(dispatch_get_main_queue(), ^{
+                handler();
+            });
+            return monitor.canConsumeGlobalKeyEvents ? NULL : event;
+        }
+
         // Any keyDown (not handled by number keys above) dismisses the overlay.
         // The event is NOT consumed — it passes through to the target app.
         if (type == kCGEventKeyDown && monitor.anyKeyDismissHandler) {
@@ -340,6 +349,12 @@ - (BOOL)handleNSEvent:(NSEvent *)event {
             return YES;
         }
 
+        // ESC (keyCode 53) during recording → cancel session.
+        if (event.type == NSEventTypeKeyDown && keyCode == 53 && self.escapeKeyHandler) {
+            self.escapeKeyHandler();
+            return YES; // consume
+        }
+
         // Some macOS versions send modifier keys as keyDown/keyUp events. Keep
         // a direct keyDown/keyUp fallback for modifier-only triggers like Fn.
         BOOL shouldHandleTriggerKeyEvent = NO;

KoeApp/Koe/Localization/SPLocalization.h

@@ -49,6 +49,14 @@ extern NSNotificationName const SPLocalizationLanguageDidChangeNotification;
 /// changes.
 + (void)invalidateCache;
 
+/// Returns every language code shipped in the app bundle (each .lproj
+/// folder containing Localizable.strings). E.g. @[@"en", @"zh-Hans", ...].
++ (NSArray<NSString *> *)availableLanguages;
+
+/// Returns a human-readable endonym for a language code (e.g. "en" →
+/// "English", "zh-Hans" → "简体中文"), suitable for a language picker.
++ (NSString *)displayNameForLanguage:(NSString *)code;
+
 @end
 
 NS_ASSUME_NONNULL_END

KoeApp/Koe/Localization/SPLocalization.m

@@ -97,4 +97,34 @@ + (NSBundle *)bundleForLanguage:(NSString *)language {
     return [NSBundle mainBundle];
 }
 
++ (NSArray<NSString *> *)availableLanguages {
+    NSFileManager *fm = [NSFileManager defaultManager];
+    NSString *resourcePath = [NSBundle mainBundle].resourcePath;
+    if (!resourcePath) return @[];
+
+    NSArray<NSString *> *entries = [fm contentsOfDirectoryAtPath:resourcePath error:NULL] ?: @[];
+    NSMutableArray<NSString *> *langs = [NSMutableArray array];
+    for (NSString *entry in entries) {
+        if (![entry.pathExtension isEqualToString:@"lproj"]) continue;
+        NSString *lproj = [resourcePath stringByAppendingPathComponent:entry];
+        NSString *strings = [lproj stringByAppendingPathComponent:@"Localizable.strings"];
+        if (![fm fileExistsAtPath:strings]) continue;
+        NSString *code = [entry stringByDeletingPathExtension];
+        [langs addObject:code];
+    }
+    [langs sortUsingSelector:@selector(caseInsensitiveCompare:)];
+    return langs;
+}
+
++ (NSString *)displayNameForLanguage:(NSString *)code {
+    if (!code.length) return @"";
+    NSLocale *targetLocale = [NSLocale localeWithLocaleIdentifier:code];
+    NSString *name = [targetLocale localizedStringForLocaleIdentifier:code];
+    if (name.length) {
+        NSString *first = [[name substringToIndex:1] uppercaseStringWithLocale:targetLocale];
+        return [name stringByReplacingCharactersInRange:NSMakeRange(0, 1) withString:first];
+    }
+    return code;
+}
+
 @end