diff --git a/__pycache__/__init__.cpython-312.pyc b/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 0000000..89df6e4
Binary files /dev/null and b/__pycache__/__init__.cpython-312.pyc differ
diff --git a/__pycache__/image.cpython-312.pyc b/__pycache__/image.cpython-312.pyc
new file mode 100644
index 0000000..45127b9
Binary files /dev/null and b/__pycache__/image.cpython-312.pyc differ
diff --git a/__pycache__/loop.cpython-312.pyc b/__pycache__/loop.cpython-312.pyc
index ae70f83..eba9e68 100644
Binary files a/__pycache__/loop.cpython-312.pyc and b/__pycache__/loop.cpython-312.pyc differ
diff --git a/__pycache__/video.cpython-312.pyc b/__pycache__/video.cpython-312.pyc
index f69fb31..36f2065 100644
Binary files a/__pycache__/video.cpython-312.pyc and b/__pycache__/video.cpython-312.pyc differ
diff --git a/image.py b/image.py
index 0dc6e49..06db9dc 100644
--- a/image.py
+++ b/image.py
@@ -159,21 +159,30 @@ class SessionImageSender:
         loop.global_accumulated_frames += advanced_original_frames
 
         source_total = getattr(loop, 'global_source_frame_count', 1)
+
+        # 🚀 NUEVO: Detectar si es el ciclo final o un ciclo único
+        is_final_cycle = loop.global_accumulated_frames >= source_total
+
         print(f"   -> ✂️ Tensor truncado a {frames_accepted} frames válidos.")
         print(f"   -> 📈 Timeline avanzado a {loop.global_accumulated_frames} / {source_total}")
 
         last_frame = valid_images[-1:].clone().cpu()
         global_session_image = last_frame
 
-        out_dir = folder_paths.get_output_directory()
-        filename = f"keyframe_{loop_idx:03d}.png"
-        filepath = os.path.join(out_dir, filename)
+        # 🛡️ Solo guardamos el keyframe si habrá un próximo ciclo
+        if not is_final_cycle:
+            out_dir = folder_paths.get_output_directory()
+            filename = f"keyframe_{loop_idx:03d}.png"
+            filepath = os.path.join(out_dir, filename)
 
-        img_array = 255. * last_frame[0].numpy()
-        img = Image.fromarray(np.clip(img_array, 0, 255).astype(np.uint8))
-        img.save(filepath)
+            img_array = 255. * last_frame[0].numpy()
+            img = Image.fromarray(np.clip(img_array, 0, 255).astype(np.uint8))
+            img.save(filepath)
+
+            print(f"   -> 💾 Keyframe seguro guardado: {filename}")
+        else:
+            print(f"   -> 🚀 Ciclo final/único detectado. Omitiendo guardado de keyframe.")
 
-        print(f"   -> 💾 Keyframe seguro guardado: {filename}")
         print(f"{'='*50}\n")
 
         ui_image = tensor_to_temp_image(last_frame, "sender")
diff --git a/video.py b/video.py
index 5e96eda..b6c857a 100644
--- a/video.py
+++ b/video.py
@@ -306,8 +306,8 @@ class IncrementalVideoStitcher:
             }
         }
 
-    RETURN_TYPES = ("IMAGE", "AUDIO")
-    RETURN_NAMES = ("ALL_IMAGES", "AUDIO_OUT")
+    RETURN_TYPES = ("IMAGE", "AUDIO", "BOOLEAN")
+    RETURN_NAMES = ("ALL_IMAGES", "AUDIO_OUT", "IS_FINAL_CYCLE")
     FUNCTION = "stitch"
     CATEGORY = "🔁 Sequential Batcher/Video"
 
@@ -322,25 +322,31 @@ class IncrementalVideoStitcher:
         torch.save(images.cpu(), path)
         print(f"🎞️ [Stitcher] Lote {current_loop_index} guardado en disco ({images.shape[0]} frames).")
 
-        is_final = loop.global_accumulated_frames >= loop.global_source_frame_count
+        source_total = getattr(loop, 'global_source_frame_count', 1)
+        is_final_cycle = loop.global_accumulated_frames >= source_total
 
-        if not is_final:
-            preview_frame = images[0:1]
-            return (preview_frame, None)
+        if is_final_cycle:
+            print(f"   -> 🏁 Generación completa. Liberando tensor total al flujo para procesos finales.")
+            print(f"📦 [Stitcher] Ensamblando todos los lotes de vídeo...")
+            all_tensors = []
+            for i in range(current_loop_index + 1):
+                p = os.path.join(cache_dir, f"batch_{i:04d}.pt")
+                if os.path.exists(p):
+                    all_tensors.append(torch.load(p))
+                    try: os.remove(p)
+                    except: pass
 
-        print(f"📦 [Stitcher] Ensamblando todos los lotes de vídeo...")
-        all_tensors = []
-        for i in range(current_loop_index + 1):
-            p = os.path.join(cache_dir, f"batch_{i:04d}.pt")
-            if os.path.exists(p):
-                all_tensors.append(torch.load(p))
-                try: os.remove(p)
-                except: pass
+            final_tensor = torch.cat(all_tensors, dim=0)
+            print(f"✅ [Stitcher] Vídeo completado: {final_tensor.shape[0]} frames ensamblados.")
 
-        final_images = torch.cat(all_tensors, dim=0)
-        print(f"✅ [Stitcher] Vídeo completado: {final_images.shape[0]} frames ensamblados.")
+            # 🚀 Retornamos True al final
+            return (final_tensor, audio, True)
+        else:
+            print(f"   -> ⏳ Ciclo intermedio. Soltando 1 frame dummy...")
+            dummy_frame = images[-1:].clone()
 
-        return (final_images, audio)
+            # 🚀 Retornamos False al final
+            return (dummy_frame, None, False)
 
 @register_node
 class AutoLoopCalculatorWan: