diff --git a/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/factory/ChatServiceFactory.java b/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/factory/ChatServiceFactory.java
index f1e88a9a..8fec93e9 100644
--- a/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/factory/ChatServiceFactory.java
+++ b/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/factory/ChatServiceFactory.java
@@ -1,6 +1,8 @@
 package org.ruoyi.chat.factory;
 
+import org.ruoyi.chat.service.chat.IChatCostService;
 import org.ruoyi.chat.service.chat.IChatService;
+import org.ruoyi.chat.service.chat.proxy.BillingChatServiceProxy;
 import org.springframework.beans.BeansException;
 import org.springframework.context.ApplicationContext;
 import org.springframework.context.ApplicationContextAware;
@@ -18,13 +20,18 @@ import java.util.concurrent.ConcurrentHashMap;
 @Component
 public class ChatServiceFactory  implements ApplicationContextAware {
     private final Map<String, IChatService> chatServiceMap = new ConcurrentHashMap<>();
+    private IChatCostService chatCostService;
 
     @Override
     public void setApplicationContext(ApplicationContext applicationContext) throws BeansException {
+        // 获取计费服务
+        this.chatCostService = applicationContext.getBean(IChatCostService.class);
+        
         // 初始化时收集所有IChatService的实现
         Map<String, IChatService> serviceMap = applicationContext.getBeansOfType(IChatService.class);
         for (IChatService service : serviceMap.values()) {
-            if (service != null) {
+            if (service != null && !isBillingProxy(service)) {
+                // 只收集非代理的原始服务
                 chatServiceMap.put(service.getCategory(), service);
             }
         }
@@ -32,12 +39,33 @@ public class ChatServiceFactory  implements ApplicationContextAware {
 
     /**
      * 根据模型类别获取对应的聊天服务实现
+     * 自动应用计费代理包装
      */
     public IChatService getChatService(String category) {
+        IChatService originalService = chatServiceMap.get(category);
+        if (originalService == null) {
+            throw new IllegalArgumentException("不支持的模型类别: " + category);
+        }
+
+        // 自动包装为计费代理
+        return new BillingChatServiceProxy(originalService, chatCostService);
+    }
+
+    /**
+     * 获取原始服务（不包装代理）
+     */
+    public IChatService getOriginalService(String category) {
         IChatService service = chatServiceMap.get(category);
         if (service == null) {
             throw new IllegalArgumentException("不支持的模型类别: " + category);
         }
         return service;
     }
+
+    /**
+     * 判断是否为计费代理实例
+     */
+    private boolean isBillingProxy(IChatService service) {
+        return service instanceof BillingChatServiceProxy;
+    }
 }
diff --git a/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/listener/SSEEventSourceListener.java b/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/listener/SSEEventSourceListener.java
index ceeb9e7b..52e94b9e 100644
--- a/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/listener/SSEEventSourceListener.java
+++ b/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/listener/SSEEventSourceListener.java
@@ -84,6 +84,8 @@ public class SSEEventSourceListener extends EventSourceListener {
                 emitter.complete();
                 // 清理失败回调（以 emitter 为键）
                 RetryNotifier.clear(emitter);
+                // 🔥 注释：AI回复的保存和计费已由BillingChatServiceProxy统一处理，此处代码已废弃
+                /*
                 // 扣除费用
                 ChatRequest chatRequest = new ChatRequest();
                 // 设置对话角色
@@ -97,6 +99,7 @@ public class SSEEventSourceListener extends EventSourceListener {
                 // 先保存助手消息，再发布异步计费事件
                 chatCostService.saveMessage(chatRequest);
                 chatCostService.publishBillingEvent(chatRequest);
+                */
                 return;
             }
 
diff --git a/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/IChatCostService.java b/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/IChatCostService.java
index 36104f9f..ccdd0617 100644
--- a/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/IChatCostService.java
+++ b/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/IChatCostService.java
@@ -61,4 +61,12 @@ public interface IChatCostService {
      * 获取登录用户id
      */
     Long getUserId();
+
+    /**
+     * 检查用户余额是否足够支付预估费用
+     * 
+     * @param chatRequest 对话信息
+     * @return true=余额充足，false=余额不足
+     */
+    boolean checkBalanceSufficient(ChatRequest chatRequest);
 }
diff --git a/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/impl/ChatCostServiceImpl.java b/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/impl/ChatCostServiceImpl.java
index 9c1cf04f..16507b94 100644
--- a/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/impl/ChatCostServiceImpl.java
+++ b/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/impl/ChatCostServiceImpl.java
@@ -428,4 +428,29 @@ public class ChatCostServiceImpl implements IChatCostService {
         }
         return loginUser.getUserId();
     }
+
+    /**
+     * 检查用户余额是否足够支付预估费用
+     */
+    @Override
+    public boolean checkBalanceSufficient(ChatRequest chatRequest) {
+        if (chatRequest.getUserId() == null) {
+            log.warn("checkBalanceSufficient->用户ID为空，视为余额不足");
+            return false;
+        }
+
+        try {
+            // 重用现有的预检查逻辑，但不抛异常，只返回boolean
+            preCheckBalance(chatRequest);
+            return true; // 预检查通过，余额充足
+        } catch (ServiceException e) {
+            log.debug("checkBalanceSufficient->余额不足，用户ID: {}, 模型: {}, 错误: {}", 
+                      chatRequest.getUserId(), chatRequest.getModel(), e.getMessage());
+            return false; // 预检查失败，余额不足
+        } catch (Exception e) {
+            log.error("checkBalanceSufficient->检查余额时发生异常，用户ID: {}, 模型: {}", 
+                      chatRequest.getUserId(), chatRequest.getModel(), e);
+            return false; // 异常情况视为余额不足，保守处理
+        }
+    }
 }
diff --git a/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/impl/DifyServiceImpl.java b/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/impl/DifyServiceImpl.java
index 66b32e6f..43608516 100644
--- a/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/impl/DifyServiceImpl.java
+++ b/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/impl/DifyServiceImpl.java
@@ -110,12 +110,12 @@ public class DifyServiceImpl implements IChatService {
                     }
 
                     // 设置对话角色
-                    chatRequestResponse.setRole(Message.Role.ASSISTANT.getName());
-                    chatRequestResponse.setModel(chatRequest.getModel());
-                    chatRequestResponse.setUserId(chatRequest.getUserId());
-                    chatRequestResponse.setSessionId(chatRequest.getSessionId());
-                    chatRequestResponse.setPrompt(respMessage.toString());
-                    chatCostService.deductToken(chatRequestResponse);
+//                    chatRequestResponse.setRole(Message.Role.ASSISTANT.getName());
+//                    chatRequestResponse.setModel(chatRequest.getModel());
+//                    chatRequestResponse.setUserId(chatRequest.getUserId());
+//                    chatRequestResponse.setSessionId(chatRequest.getSessionId());
+//                    chatRequestResponse.setPrompt(respMessage.toString());
+//                    chatCostService.deductToken(chatRequestResponse);
                     RetryNotifier.clear(emitter);
                 }
 
diff --git a/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/impl/SseServiceImpl.java b/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/impl/SseServiceImpl.java
index eb9dba8b..0250177a 100644
--- a/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/impl/SseServiceImpl.java
+++ b/ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/chat/service/chat/impl/SseServiceImpl.java
@@ -116,8 +116,6 @@ public class SseServiceImpl implements ISseService {
                 }
 
 
-                // 先保存消息，再发布异步计费事件
-                chatCostService.saveMessage(chatRequest);
 
                 chatRequest.setUserId(chatCostService.getUserId());
                 if (chatRequest.getSessionId() == null) {
@@ -128,11 +126,15 @@ public class SseServiceImpl implements ISseService {
                     chatSessionService.insertByBo(chatSessionBo);
                     chatRequest.setSessionId(chatSessionBo.getId());
                 }
+                
+                // 保存用户消息
+                chatCostService.saveMessage(chatRequest);
             }
             // 自动选择模型并获取对应的聊天服务
             IChatService chatService = autoSelectModelAndGetService(chatRequest);
-            chatCostService.publishBillingEvent(chatRequest);
-            // 仅当 autoSelectModel = true 时，才启用重试与降级
+
+            // 用户消息只保存不计费，AI回复由BillingChatServiceProxy自动处理计费
+            // chatCostService.publishBillingEvent(chatRequest); // 用户输入不计费
             if (Boolean.TRUE.equals(chatRequest.getAutoSelectModel())) {
                 ChatModelVo currentModel = this.chatModelVo;
                 String currentCategory = currentModel.getCategory();