Fix collision stuck state + add streaming partial text event

- Skip mic buffer flush during collision avoidance (bAgentGenerating guard
  in StopListening) to prevent sending audio to a mid-generation server
  which caused both sides to stall permanently
- Add OnAgentPartialResponse event: streams LLM text fragments from
  agent_chat_response_part in real-time (opt-in via bEnableAgentPartialResponse),
  separate from the existing OnAgentTextResponse (full text at end)
- French agent server drop after 3 turns is a server-side issue, not client

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Unreal/PS_AI_Agent/Plugins/PS_AI_Agent_ElevenLabs/Source/PS_AI_Agent_ElevenLabs/Private/ElevenLabsConversationalAgentComponent.cpp

@@ -158,6 +158,8 @@ void UElevenLabsConversationalAgentComponent::StartConversation()
 			&UElevenLabsConversationalAgentComponent::HandleInterrupted);
 		WebSocketProxy->OnAgentResponseStarted.AddDynamic(this,
 			&UElevenLabsConversationalAgentComponent::HandleAgentResponseStarted);
+		WebSocketProxy->OnAgentResponsePart.AddDynamic(this,
+			&UElevenLabsConversationalAgentComponent::HandleAgentResponsePart);
 	}
 
 	// Pass configuration to the proxy before connecting.
@@ -266,7 +268,23 @@ void UElevenLabsConversationalAgentComponent::StopListening()
 	// Flush any partially-accumulated mic audio before signalling end-of-turn.
 	// This ensures the final words aren't discarded just because the last callback
 	// didn't push the buffer over the MicChunkMinBytes threshold.
-	if (MicAccumulationBuffer.Num() > 0 && WebSocketProxy && IsConnected())
+	//
+	// EXCEPT during collision avoidance: bAgentGenerating is already true when
+	// HandleAgentResponseStarted calls StopListening (it sets the flag before calling us).
+	// Flushing audio to a server that is mid-generation can cause it to re-enter
+	// "user speaking" state and stall waiting for more audio that never arrives,
+	// leaving both sides stuck — no audio for the collision response and no response
+	// for subsequent turns.
+	if (bAgentGenerating)
+	{
+		if (MicAccumulationBuffer.Num() > 0)
+		{
+			UE_LOG(LogElevenLabsAgent, Log,
+				TEXT("StopListening: discarding %d bytes of accumulated mic audio (collision — server is mid-generation)."),
+				MicAccumulationBuffer.Num());
+		}
+	}
+	else if (MicAccumulationBuffer.Num() > 0 && WebSocketProxy && IsConnected())
 	{
 		WebSocketProxy->SendAudioChunk(MicAccumulationBuffer);
 	}
@@ -423,7 +441,8 @@ void UElevenLabsConversationalAgentComponent::HandleAgentResponseStarted()
 {
 	// The server has started generating a response (first agent_chat_response_part).
 	// Set bAgentGenerating BEFORE StopListening so that any StartListening call
-	// triggered by the Blueprint's OnAgentStartedGenerating handler is blocked.
+	// triggered by the Blueprint's OnAgentStartedGenerating handler is blocked,
+	// and so that StopListening knows to skip the mic buffer flush (collision path).
 	bAgentGenerating = true;
 	bWaitingForAgentResponse = false;  // Server is generating — response timeout cancelled.
 
@@ -433,21 +452,39 @@ void UElevenLabsConversationalAgentComponent::HandleAgentResponseStarted()
 	if (bIsListening)
 	{
 		// Collision: server started generating Turn N's response while Turn M (M>N) mic was open.
-		// Log both turn indices so the timeline is unambiguous.
+		// The server's VAD detected a pause in the user's speech and started generating
+		// prematurely — the user hasn't finished speaking yet.
+		//
+		// Stop the mic WITHOUT flushing the accumulated audio buffer (see StopListening's
+		// bAgentGenerating guard). Flushing would send audio to a server that is mid-generation,
+		// causing it to re-enter "user speaking" state and stall — both sides stuck.
+		//
+		// Do NOT send an interrupt here: the ElevenLabs server does not always send the
+		// interruption ack, which would leave bIgnoreIncomingContent=true and silently
+		// discard all subsequent content. Instead, let the server's response play out:
+		//   - If audio arrives → EnqueueAgentAudio sets bAgentSpeaking, response plays normally.
+		//   - If audio never arrives → generating timeout (10s) clears bAgentGenerating.
+		// Either way the state machine recovers and Blueprint can reopen the mic.
 		UE_LOG(LogElevenLabsAgent, Log,
 			TEXT("[T+%.2fs] [Turn %d → Turn %d collision] Agent generating Turn %d response — mic (Turn %d) was open, stopping. (%.2fs after turn end)"),
 			T, LastClosedTurnIndex, TurnIndex, LastClosedTurnIndex, TurnIndex, LatencyFromTurnEnd);
 		StopListening();
 	}
-	else
-	{
-		UE_LOG(LogElevenLabsAgent, Log,
-			TEXT("[T+%.2fs] [Turn %d] Agent generating. (%.2fs after turn end)"),
-			T, LastClosedTurnIndex, LatencyFromTurnEnd);
-	}
+
+	UE_LOG(LogElevenLabsAgent, Log,
+		TEXT("[T+%.2fs] [Turn %d] Agent generating. (%.2fs after turn end)"),
+		T, LastClosedTurnIndex, LatencyFromTurnEnd);
 	OnAgentStartedGenerating.Broadcast();
 }
 
+void UElevenLabsConversationalAgentComponent::HandleAgentResponsePart(const FString& PartialText)
+{
+	if (bEnableAgentPartialResponse)
+	{
+		OnAgentPartialResponse.Broadcast(PartialText);
+	}
+}
+
 // ─────────────────────────────────────────────────────────────────────────────
 // Audio playback
 // ─────────────────────────────────────────────────────────────────────────────

Unreal/PS_AI_Agent/Plugins/PS_AI_Agent_ElevenLabs/Source/PS_AI_Agent_ElevenLabs/Private/ElevenLabsWebSocketProxy.cpp

@@ -400,7 +400,7 @@ void UElevenLabsWebSocketProxy::OnWsMessage(const FString& Message)
 	}
 	else if (MsgType == ElevenLabsMessageType::AgentChatResponsePart)
 	{
-		HandleAgentChatResponsePart();
+		HandleAgentChatResponsePart(Root);
 	}
 	else if (MsgType == ElevenLabsMessageType::AgentResponseCorrection)
 	{
@@ -602,7 +602,7 @@ void UElevenLabsWebSocketProxy::HandleAgentResponse(const TSharedPtr<FJsonObject
 	OnAgentResponse.Broadcast(ResponseText);
 }
 
-void UElevenLabsWebSocketProxy::HandleAgentChatResponsePart()
+void UElevenLabsWebSocketProxy::HandleAgentChatResponsePart(const TSharedPtr<FJsonObject>& Root)
 {
 	// Ignore response parts that belong to a generation we have already interrupted.
 	// Without this guard, old parts arriving after SendInterrupt() would re-trigger
@@ -628,7 +628,21 @@ void UElevenLabsWebSocketProxy::HandleAgentChatResponsePart()
 			T, LatencyFromTurnEnd);
 		OnAgentResponseStarted.Broadcast();
 	}
-	// Subsequent parts logged at Verbose only (can be dozens per response).
+
+	// Extract the streaming text fragment and broadcast it.
+	// API structure:
+	// { "type": "agent_chat_response_part",
+	//   "agent_chat_response_part_event": { "agent_response_part": "partial text" }
+	// }
+	const TSharedPtr<FJsonObject>* PartEvent = nullptr;
+	if (Root->TryGetObjectField(TEXT("agent_chat_response_part_event"), PartEvent) && PartEvent)
+	{
+		FString PartText;
+		if ((*PartEvent)->TryGetStringField(TEXT("agent_response_part"), PartText) && !PartText.IsEmpty())
+		{
+			OnAgentResponsePart.Broadcast(PartText);
+		}
+	}
 }
 
 void UElevenLabsWebSocketProxy::HandleInterruption(const TSharedPtr<FJsonObject>& Root)

Unreal/PS_AI_Agent/Plugins/PS_AI_Agent_ElevenLabs/Source/PS_AI_Agent_ElevenLabs/Public/ElevenLabsConversationalAgentComponent.h

@@ -43,6 +43,15 @@ DECLARE_DYNAMIC_MULTICAST_DELEGATE(FOnAgentInterrupted);
  */
 DECLARE_DYNAMIC_MULTICAST_DELEGATE(FOnAgentStartedGenerating);
 
+/**
+ * Fired for every agent_chat_response_part — streams the agent's text as the LLM
+ * generates it, token by token. Use this for real-time subtitles / text display.
+ * Each call provides the text fragment from that individual part (NOT accumulated).
+ * The final complete text is still available via OnAgentTextResponse (agent_response).
+ */
+DECLARE_DYNAMIC_MULTICAST_DELEGATE_OneParam(FOnAgentPartialResponse,
+	const FString&, PartialText);
+
 /**
  * Fired when the server has not started generating a response within ResponseTimeoutSeconds
  * after the user stopped speaking (StopListening was called).
@@ -138,6 +147,15 @@ public:
 	UPROPERTY(EditAnywhere, BlueprintReadWrite, Category = "ElevenLabs|Events")
 	bool bEnableAgentTextResponse = true;
 
+	/**
+	 * Forward streaming text parts (agent_chat_response_part events) to the
+	 * OnAgentPartialResponse delegate. Each part is a text fragment as the LLM
+	 * generates it — use this for real-time subtitles that appear while the agent
+	 * speaks, instead of waiting for the full text (OnAgentTextResponse).
+	 */
+	UPROPERTY(EditAnywhere, BlueprintReadWrite, Category = "ElevenLabs|Events")
+	bool bEnableAgentPartialResponse = false;
+
 	/**
 	 * How many seconds to wait for the server to start generating a response
 	 * after the user stops speaking (StopListening) before firing OnAgentResponseTimeout.
@@ -168,6 +186,14 @@ public:
 	UPROPERTY(BlueprintAssignable, Category = "ElevenLabs|Events")
 	FOnAgentTextResponse OnAgentTextResponse;
 
+	/**
+	 * Streaming text fragments as the LLM generates them.
+	 * Fires for every agent_chat_response_part — each call gives one text chunk.
+	 * Enable with bEnableAgentPartialResponse.
+	 */
+	UPROPERTY(BlueprintAssignable, Category = "ElevenLabs|Events")
+	FOnAgentPartialResponse OnAgentPartialResponse;
+
 	UPROPERTY(BlueprintAssignable, Category = "ElevenLabs|Events")
 	FOnAgentStartedSpeaking OnAgentStartedSpeaking;
 
@@ -285,6 +311,9 @@ private:
 	UFUNCTION()
 	void HandleAgentResponseStarted();
 
+	UFUNCTION()
+	void HandleAgentResponsePart(const FString& PartialText);
+
 	// ── Audio playback ────────────────────────────────────────────────────────
 	void InitAudioPlayback();
 	void EnqueueAgentAudio(const TArray<uint8>& PCMData);

Unreal/PS_AI_Agent/Plugins/PS_AI_Agent_ElevenLabs/Source/PS_AI_Agent_ElevenLabs/Public/ElevenLabsWebSocketProxy.h

@@ -43,6 +43,11 @@ DECLARE_DYNAMIC_MULTICAST_DELEGATE(FOnElevenLabsInterrupted);
  */
 DECLARE_DYNAMIC_MULTICAST_DELEGATE(FOnElevenLabsAgentResponseStarted);
 
+/** Fired for every agent_chat_response_part — streams the LLM text as it is generated.
+ *  PartialText is the text fragment from this individual part (NOT accumulated). */
+DECLARE_DYNAMIC_MULTICAST_DELEGATE_OneParam(FOnElevenLabsAgentResponsePart,
+	const FString&, PartialText);
+
 
 // ─────────────────────────────────────────────────────────────────────────────
 // WebSocket Proxy
@@ -94,6 +99,10 @@ public:
 	UPROPERTY(BlueprintAssignable, Category = "ElevenLabs|Events")
 	FOnElevenLabsAgentResponseStarted OnAgentResponseStarted;
 
+	/** Fired for every agent_chat_response_part with the streaming text fragment. */
+	UPROPERTY(BlueprintAssignable, Category = "ElevenLabs|Events")
+	FOnElevenLabsAgentResponsePart OnAgentResponsePart;
+
 	// ── Lifecycle ─────────────────────────────────────────────────────────────
 
 	/**
@@ -182,7 +191,7 @@ private:
 	void HandleAudioResponse(const TSharedPtr<FJsonObject>& Payload);
 	void HandleTranscript(const TSharedPtr<FJsonObject>& Payload);
 	void HandleAgentResponse(const TSharedPtr<FJsonObject>& Payload);
-	void HandleAgentChatResponsePart();
+	void HandleAgentChatResponsePart(const TSharedPtr<FJsonObject>& Payload);
 	void HandleInterruption(const TSharedPtr<FJsonObject>& Payload);
 	void HandlePing(const TSharedPtr<FJsonObject>& Payload);