diff --git a/frontend/src/components/QRCodeCard.vue b/frontend/src/components/QRCodeCard.vue
index cbe0a9c..5807185 100644
--- a/frontend/src/components/QRCodeCard.vue
+++ b/frontend/src/components/QRCodeCard.vue
@@ -1,5 +1,5 @@
 <template>
-    <div class="flex flex-col items-center space-y-4 p-6 bg-base-100 rounded-xl shadow-lg max-w-sm mx-auto backdrop-blur-xl glass">
+    <div class="flex flex-col items-center space-y-4 p-6 bg-base-100 rounded-xl shadow-lg mt-2 max-w-sm mx-auto backdrop-blur-xl glass">
 
         <div class="p-3 bg-white rounded-lg shadow-inner cursor-pointer" @click="toggleQRCode">
             <qrcode-vue :value="currentValue" :size="size" level="H" />
diff --git a/internal/controller/proxy/chat_proxy.go b/internal/controller/proxy/chat_proxy.go
index 96851c3..abbc78b 100644
--- a/internal/controller/proxy/chat_proxy.go
+++ b/internal/controller/proxy/chat_proxy.go
@@ -60,19 +60,22 @@ func (h *Proxy) ChatHandler(c *gin.Context) {
 			c.SSEvent("", data)
 		}
 	}
-	llmusage := llm.GetTokenUsage()
+	go func() {
+		llmusage := llm.GetTokenUsage()
+
+		cost := tokenizer.Cost(llmusage.Model, llmusage.PromptTokens+llmusage.ToolsTokens, llmusage.CompletionTokens)
+		userid, _ := strconv.ParseInt(c.GetString("user_id"), 10, 64)
+		usage := model.Usage{
+			UserID:           userid,
+			Model:            llmusage.Model,
+			Stream:           chatreq.Stream,
+			PromptTokens:     llmusage.PromptTokens + llmusage.ToolsTokens,
+			CompletionTokens: llmusage.CompletionTokens,
+			TotalTokens:      llmusage.TotalTokens,
+			Cost:             fmt.Sprintf("%f", cost),
+		}
+		h.SendUsage(&usage)
+		defer fmt.Println("cost:", cost, "prompt_tokens:", llmusage.PromptTokens, "completion_tokens:", llmusage.CompletionTokens, "total_tokens:", llmusage.TotalTokens)
+	}()
 
-	cost := tokenizer.Cost(llmusage.Model, llmusage.PromptTokens+llmusage.ToolsTokens, llmusage.CompletionTokens)
-	userid, _ := strconv.ParseInt(c.GetString("user_id"), 10, 64)
-	usage := model.Usage{
-		UserID:           userid,
-		Model:            llmusage.Model,
-		Stream:           chatreq.Stream,
-		PromptTokens:     llmusage.PromptTokens + llmusage.ToolsTokens,
-		CompletionTokens: llmusage.CompletionTokens,
-		TotalTokens:      llmusage.TotalTokens,
-		Cost:             fmt.Sprintf("%f", cost),
-	}
-	h.SendUsage(&usage)
-	defer fmt.Println("cost:", cost, "prompt_tokens:", llmusage.PromptTokens, "completion_tokens:", llmusage.CompletionTokens, "total_tokens:", llmusage.TotalTokens)
 }