diff --git a/vision_agent/tools/tools.py b/vision_agent/tools/tools.py
index 898197a9..5f3ca93e 100644
--- a/vision_agent/tools/tools.py
+++ b/vision_agent/tools/tools.py
@@ -467,10 +467,22 @@ def florence2_sam2_video_tracking(
         return_data.append(return_frame_data)
     return_data = add_bboxes_from_masks(return_data)
     return_data = nms(return_data, iou_threshold=0.95)
+
     _display_tool_trace(
         florence2_sam2_video_tracking.__name__,
         payload,
-        detections[0],
+        [
+            [
+                {
+                    "label": e["label"],
+                    "score": e["score"],
+                    "bbox": denormalize_bbox(e["bbox"], frames[0].shape[:2]),
+                    "mask": rle_encode_array(e["mask"]),
+                }
+                for e in lst
+            ]
+            for lst in return_data
+        ],
         files,
     )
     return return_data
@@ -686,6 +698,11 @@ def _run_countgd(prompt: str) -> List[Dict[str, Any]]:
         )
         for bbox in bboxes
     ]
+
+    # TODO: remove this once we start to use the confidence on countgd
+    filtered_bboxes = filter_bboxes_by_threshold(bboxes_formatted, box_threshold)
+    return_data = [bbox.model_dump() for bbox in filtered_bboxes]
+    return_data = single_nms(return_data, iou_threshold=0.80)
     _display_tool_trace(
         countgd_object_detection.__name__,
         {
@@ -693,14 +710,17 @@ def _run_countgd(prompt: str) -> List[Dict[str, Any]]:
             "confidence": box_threshold,
             "model": "countgd",
         },
-        bboxes,
+        [
+            {
+                "label": e["label"],
+                "score": e["score"],
+                "bbox": denormalize_bbox(e["bbox"], image_size),
+            }
+            for e in return_data
+        ],
         files,
     )
-
-    # TODO: remove this once we start to use the confidence on countgd
-    filtered_bboxes = filter_bboxes_by_threshold(bboxes_formatted, box_threshold)
-    return_data = [bbox.model_dump() for bbox in filtered_bboxes]
-    return single_nms(return_data, iou_threshold=0.80)
+    return return_data
 
 
 def countgd_sam2_object_detection(
@@ -830,14 +850,21 @@ def countgd_example_based_counting(
         )
         for bbox in bboxes_per_frame
     ]
+    filtered_bboxes = filter_bboxes_by_threshold(bboxes_formatted, box_threshold)
     _display_tool_trace(
         countgd_example_based_counting.__name__,
         payload,
-        detections[0],
+        [
+            {
+                "label": e.label,
+                "score": e.score,
+                "bbox": denormalize_bbox(e.bbox, image_size),  # type: ignore
+            }
+            for e in filtered_bboxes
+        ],
         files,
     )
 
-    filtered_bboxes = filter_bboxes_by_threshold(bboxes_formatted, box_threshold)
     return [bbox.model_dump() for bbox in filtered_bboxes]