diff --git a/esphome/components/voice_assistant/voice_assistant.h b/esphome/components/voice_assistant/voice_assistant.h
index d6b1502381..1c0ea12f4f 100644
--- a/esphome/components/voice_assistant/voice_assistant.h
+++ b/esphome/components/voice_assistant/voice_assistant.h
@@ -94,10 +94,10 @@ class VoiceAssistant : public Component {
   uint32_t get_feature_flags() const {
     uint32_t flags = 0;
     flags |= VoiceAssistantFeature::FEATURE_VOICE_ASSISTANT;
+    flags |= VoiceAssistantFeature::FEATURE_API_AUDIO;
 #ifdef USE_SPEAKER
     if (this->speaker_ != nullptr) {
       flags |= VoiceAssistantFeature::FEATURE_SPEAKER;
-      flags |= VoiceAssistantFeature::FEATURE_API_AUDIO;
     }
 #endif
     return flags;