diff --git a/main/application.cc b/main/application.cc
index 981a675..1095710 100644
--- a/main/application.cc
+++ b/main/application.cc
@@ -682,6 +682,10 @@ void Application::ToggleChatStateWithVision() {
     xEventGroupSetBits(event_group_, MAIN_EVENT_TOGGLE_CHAT);
 }
 
+bool Application::IsVisionTextModeEnabled() const {
+    return vision_text_mode_enabled_.load();
+}
+
 void Application::StartListening() {
     vision_text_mode_enabled_.store(false);
     xEventGroupSetBits(event_group_, MAIN_EVENT_START_LISTENING);
@@ -968,7 +972,7 @@ void Application::SendCurrentVisionFrame() {
     }
 
     std::string jpeg_data;
-    if (!camera->CaptureToJpeg(jpeg_data, false)) {
+    if (!camera->CaptureToJpeg(jpeg_data, true)) {
         ESP_LOGW(TAG, "Failed to capture vision frame");
         return;
     }
diff --git a/main/application.h b/main/application.h
index 8aae2c5..a4fb7a5 100644
--- a/main/application.h
+++ b/main/application.h
@@ -93,6 +93,7 @@ public:
      */
     void ToggleChatState();
     void ToggleChatStateWithVision();
+    bool IsVisionTextModeEnabled() const;
 
     /**
      * Start listening (event-based, thread-safe)
diff --git a/main/boards/m5stack-core-s3/m5stack_core_s3.cc b/main/boards/m5stack-core-s3/m5stack_core_s3.cc
index 0513491..33c8cd4 100644
--- a/main/boards/m5stack-core-s3/m5stack_core_s3.cc
+++ b/main/boards/m5stack-core-s3/m5stack_core_s3.cc
@@ -349,7 +349,10 @@ private:
                         continue;
                     }
 
-                    if (board->camera_->CaptureBackground()) {
+                    bool capture_ok = Application::GetInstance().IsVisionTextModeEnabled()
+                                          ? board->camera_->Capture()
+                                          : board->camera_->CaptureBackground();
+                    if (capture_ok) {
                         if (!has_logged_success) {
                             ESP_LOGI(TAG, "Background vision sampler started");
                             has_logged_success = true;