diff --git a/scripts/serve-frank-large.sh b/scripts/serve-frank-large.sh
index 0c518fa..6daf828 100755
--- a/scripts/serve-frank-large.sh
+++ b/scripts/serve-frank-large.sh
@@ -81,6 +81,7 @@ ARGS=(
     --parallel "$PARALLEL_SLOTS"
     --n-gpu-layers 999
     --jinja
+    --metrics
     --chat-template-file "$TEMPLATE_FILE"
     --chat-template-kwargs '{"preserve_thinking":true}'
     --cache-type-k q8_0
diff --git a/scripts/serve-frank-primary.sh b/scripts/serve-frank-primary.sh
index 6fd1f4d..d91748d 100755
--- a/scripts/serve-frank-primary.sh
+++ b/scripts/serve-frank-primary.sh
@@ -56,6 +56,7 @@ fi
 : "${BATCH_SIZE:=8192}"
 : "${UBATCH_SIZE:=512}"
 : "${CACHE_RAM:=16384}"
+: "${VERBOSITY:=4}"
 
 SERVER_BIN="$RUNTIME_DIR/llama-server"
 MODEL_PARTS=(
@@ -99,6 +100,8 @@ ARGS=(
     --alias frank-primary
     --ctx-size "$CONTEXT_SIZE"
     --parallel "$PARALLEL_SLOTS"
+    -lv "$VERBOSITY"
+    --metrics
     --n-gpu-layers "$NGL"
     --tensor-split "$TENSOR_SPLIT"
     --jinja