diff --git a/.env.template b/.env.template
index 84d8ba84..ebe7fa16 100644
--- a/.env.template
+++ b/.env.template
@@ -8,6 +8,7 @@
 # Application Settings
 # =============================================================================
 LOG_LEVEL=INFO
+PERFORMANCE_LOG_FORMAT=compact # compact|rich
 # SESSION_OBSERVERS_LIMIT=10
 # GET_CONTEXT_MAX_TOKENS=100000
 # MAX_FILE_SIZE=5242880 # Bytes
diff --git a/config.toml.example b/config.toml.example
index 9e0e9c11..54bf1ffd 100644
--- a/config.toml.example
+++ b/config.toml.example
@@ -6,6 +6,7 @@
 # Application-level settings
 [app]
 LOG_LEVEL = "INFO"
+PERFORMANCE_LOG_FORMAT = "compact" # "compact" for single-line logs, "rich" for local panels
 SESSION_OBSERVERS_LIMIT = 10
 GET_CONTEXT_MAX_TOKENS = 100000
 MAX_FILE_SIZE = 5242880  # 5MB
diff --git a/src/config.py b/src/config.py
index fa700d6d..5a0e3937 100644
--- a/src/config.py
+++ b/src/config.py
@@ -1353,6 +1353,7 @@ class AppSettings(HonchoSettings):
 
     # Application-wide settings
     LOG_LEVEL: str = "INFO"
+    PERFORMANCE_LOG_FORMAT: str = "compact"
     SESSION_OBSERVERS_LIMIT: Annotated[int, Field(default=10, gt=0)] = 10
     MAX_FILE_SIZE: Annotated[int, Field(default=5_242_880, gt=0)] = 5_242_880  # 5MB
     GET_CONTEXT_MAX_TOKENS: Annotated[int, Field(default=100_000, gt=0, le=250_000)] = (
@@ -1401,6 +1402,13 @@ class AppSettings(HonchoSettings):
             raise ValueError(f"Invalid log level: {v}")
         return log_level
 
+    @field_validator("PERFORMANCE_LOG_FORMAT")
+    def validate_performance_log_format(cls, v: str) -> str:
+        log_format = v.lower()
+        if log_format not in ["compact", "rich"]:
+            raise ValueError(f"Invalid performance log format: {v}")
+        return log_format
+
     @model_validator(mode="after")
     def propagate_namespace(self) -> "AppSettings":
         """Propagate top-level NAMESPACE to nested settings if not explicitly set."""
diff --git a/src/crud/document.py b/src/crud/document.py
index 018f3411..11eb121d 100644
--- a/src/crud/document.py
+++ b/src/crud/document.py
@@ -1024,8 +1024,10 @@ async def is_rejected_duplicate(
 
     # If new document has more or equal information, keep it and delete existing
     if score_new >= score_existing:
-        logger.warning(
-            f"[DUPLICATE DETECTION] Deleting existing in favor of new. new='{doc.content}', existing='{existing_doc.content}'."
+        logger.debug(
+            "[DUPLICATE DETECTION] Deleting existing in favor of new. new=%r, existing=%r.",
+            doc.content,
+            existing_doc.content,
         )
         # Carry the reinforcement count forward so replacing a duplicate counts as
         # another derivation rather than resetting times_derived to 1.
@@ -1041,8 +1043,10 @@ async def is_rejected_duplicate(
     # reinforcing the same document must not lose updates.
     existing_doc.times_derived = models.Document.times_derived + 1
     await db.flush()
-    logger.warning(
-        f"[DUPLICATE DETECTION] Rejecting new in favor of existing. new='{doc.content}', existing='{existing_doc.content}'."
+    logger.debug(
+        "[DUPLICATE DETECTION] Rejecting new in favor of existing. new=%r, existing=%r.",
+        doc.content,
+        existing_doc.content,
     )
     return True
 
diff --git a/src/deriver/enqueue.py b/src/deriver/enqueue.py
index aaeda415..05883306 100644
--- a/src/deriver/enqueue.py
+++ b/src/deriver/enqueue.py
@@ -495,7 +495,7 @@ async def enqueue_dream(
             is_in_progress = await db_session.scalar(in_progress_check)
 
             if is_in_progress:
-                logger.info(
+                logger.debug(
                     "Skipping dream enqueue - already in progress: %s/%s/%s (type: %s)",
                     workspace_name,
                     observer,
@@ -515,7 +515,7 @@ async def enqueue_dream(
             is_pending = await db_session.scalar(pending_check)
 
             if is_pending:
-                logger.info(
+                logger.debug(
                     "Dream already pending in queue: %s/%s/%s (type: %s)",
                     workspace_name,
                     observer,
diff --git a/src/dreamer/dream_scheduler.py b/src/dreamer/dream_scheduler.py
index bad850f4..43bba428 100644
--- a/src/dreamer/dream_scheduler.py
+++ b/src/dreamer/dream_scheduler.py
@@ -218,7 +218,7 @@ class DreamScheduler:
             configuration = get_configuration(None, session, workspace)
 
             if not configuration.dream.enabled:
-                logger.info(
+                logger.debug(
                     f"Dreams disabled for {workspace_name}/{session_name}, skipping dream"
                 )
                 return
@@ -319,7 +319,7 @@ async def check_and_schedule_dream(
                 ).total_seconds() / 3600
 
                 if hours_since_last_dream < settings.DREAM.MIN_HOURS_BETWEEN_DREAMS:
-                    logger.info(
+                    logger.debug(
                         f"Skipping dream for {collection.observer}/{collection.observed}: only {hours_since_last_dream:.1f} hours "
                         + f"since last dream (minimum: {settings.DREAM.MIN_HOURS_BETWEEN_DREAMS})"
                     )
@@ -359,7 +359,7 @@ async def check_and_schedule_dream(
             )
         )
         if pending_exists:
-            logger.info(
+            logger.debug(
                 "Skipping dream schedule for %s/%s: pending dream already in queue",
                 collection.observer,
                 collection.observed,
diff --git a/src/reconciler/scheduler.py b/src/reconciler/scheduler.py
index d4f314a3..e171c4fa 100644
--- a/src/reconciler/scheduler.py
+++ b/src/reconciler/scheduler.py
@@ -264,5 +264,5 @@ class ReconcilerScheduler:
                 )
                 return False
 
-            logger.info("Enqueued reconciler task: %s", task.name)
+            logger.debug("Enqueued reconciler task: %s", task.name)
             return True
diff --git a/src/reconciler/sync_vectors.py b/src/reconciler/sync_vectors.py
index cdc18542..68e9ac59 100644
--- a/src/reconciler/sync_vectors.py
+++ b/src/reconciler/sync_vectors.py
@@ -728,7 +728,7 @@ async def run_vector_reconciliation_cycle() -> ReconciliationMetrics:
 
             if not (embs_work or cleanup_work):
                 break
-        logger.info("Vector reconciliation cycle completed (pgvector mode)")
+        logger.debug("Vector reconciliation cycle completed (pgvector mode)")
         return metrics
 
     # External vector store mode - reconcile documents, embeddings, and cleanup
@@ -755,5 +755,5 @@ async def run_vector_reconciliation_cycle() -> ReconciliationMetrics:
             logger.debug("No work done, breaking reconciliation loop")
             break
 
-    logger.info("Vector reconciliation cycle completed")
+    logger.debug("Vector reconciliation cycle completed")
     return metrics
diff --git a/src/telemetry/logging.py b/src/telemetry/logging.py
index c2fcba69..441b1f96 100644
--- a/src/telemetry/logging.py
+++ b/src/telemetry/logging.py
@@ -17,7 +17,6 @@ from rich.console import Console, Group, RenderableType
 from rich.panel import Panel
 from rich.table import Table
 from rich.text import Text
-from rich.tree import Tree
 
 from src.config import settings
 from src.telemetry.metrics_collector import append_metrics_to_file
@@ -176,28 +175,6 @@ def format_reasoning_inputs_as_markdown(
     return "\n".join(parts)
 
 
-def log_representation(
-    representation: Representation,
-) -> None:
-    """
-    Log representation in a tree structure.
-    Args:
-        representation: Representation to log
-    """
-    tree = Tree("📊 REPRESENTATION")
-
-    type_branch = tree.add(f"[bold cyan]EXPLICIT[/] ({len(representation.explicit)})")
-    for i, obs in enumerate(representation.explicit, 1):
-        type_branch.add(f"[dim]{i}.[/] {obs}")
-
-    type_branch = tree.add(f"[bold cyan]DEDUCTIVE[/] ({len(representation.deductive)})")
-    for i, obs in enumerate(representation.deductive, 1):
-        type_branch.add(f"[dim]{i}.[/] {obs}")
-
-    console.print(tree)
-    console.print()
-
-
 def accumulate_metric(
     task_name: str,
     label: str,
@@ -276,16 +253,20 @@ def log_performance_metrics(
     task_slug: str,
     task_name: str,
     metrics: list[tuple[str, str | int | float, str]] | None = None,
-    title: str = "⚡ PERFORMANCE",
+    title: str = "PERFORMANCE",
 ) -> None:
     """
-    Log performance metrics in a clean table and optionally send to global collector.
+    Log performance metrics and optionally send them to the global collector.
+
+    PERFORMANCE_LOG_FORMAT=compact emits numeric metrics on one INFO line and
+    keeps large "blob" metrics at DEBUG. PERFORMANCE_LOG_FORMAT=rich prints the
+    local Rich panel, including blob metrics, for interactive readability.
 
     Args:
         task_slug: Slug of the task that generated these metrics
         task_name: Name of the task that generated these metrics
-        metrics: Dictionary of metric names and (value, unit) tuples
-        title: Table title
+        metrics: List of (metric_name, value, unit) tuples
+        title: Prefix for the log line
     """
     task_name = f"{task_slug}_{task_name}"
     # No-op if metrics were evicted (due to MAX_ACCUMULATED_TASKS limit) and no
@@ -300,7 +281,41 @@ def log_performance_metrics(
     if COLLECT_METRICS_LOCAL:
         append_metrics_to_file(task_slug, task_name, metrics)
 
-    # Remove metrics with "blob" unit type. They get printed separately below the table.
+    if settings.PERFORMANCE_LOG_FORMAT == "rich":
+        _log_performance_metrics_rich(task_name, metrics, title)
+        return
+
+    # Keep large text payloads out of the compact INFO summary.
+    blob_metrics: list[tuple[str, str | int | float, str]] = []
+    summary_parts: list[str] = []
+    for metric, value, unit in metrics:
+        if unit == "blob":
+            blob_metrics.append((metric, value, unit))
+            continue
+        if unit == "ms" and isinstance(value, int | float):
+            formatted_value = f"{value:.0f}ms"
+        elif unit == "s" and isinstance(value, int | float):
+            formatted_value = f"{value:.3f}s"
+        elif unit in ("", "tokens", "count", "id"):
+            formatted_value = str(value)
+        else:
+            formatted_value = f"{value}{unit}"
+        summary_parts.append(f"{metric}={formatted_value}")
+
+    if summary_parts:
+        logger.info("%s %s | %s", title, task_name, " | ".join(summary_parts))
+    else:
+        logger.info("%s %s", title, task_name)
+
+    for metric, value, _unit in blob_metrics:
+        logger.debug("%s %s :: %s\n%s", title, task_name, metric, value)
+
+
+def _log_performance_metrics_rich(
+    task_name: str,
+    metrics: list[tuple[str, str | int | float, str]],
+    title: str,
+) -> None:
     blob_metrics: list[tuple[str, str | int | float, str]] = []
     non_blob_metrics: list[tuple[str, str | int | float, str]] = []
     for metric in metrics:
@@ -317,24 +332,20 @@ def log_performance_metrics(
     table.add_column("Unit", style="dim", width=8)
 
     for metric, value, unit in non_blob_metrics:
-        if unit == "ms":
+        if unit == "ms" and isinstance(value, int | float):
             formatted_value = f"{value:.0f}"
-        elif unit == "s":
+        elif unit == "s" and isinstance(value, int | float):
             formatted_value = f"{value:.3f}"
         else:
             formatted_value = str(value)
 
         table.add_row(metric.replace("_", " ").title(), formatted_value, unit)
 
-    # Build content for the panel
     content_items: list[RenderableType] = [table]
 
-    if blob_metrics:
-        for metric, value, _unit in blob_metrics:
-            content_items.append(
-                Text.assemble("    ", (f"\n{metric}:", "bold"), "     ")
-            )
-            content_items.append(Text(str(value)))
+    for metric, value, _unit in blob_metrics:
+        content_items.append(Text.assemble("    ", (f"\n{metric}:", "bold"), "     "))
+        content_items.append(Text(str(value)))
 
     panel = Panel(
         Group(*content_items),