diff --git a/core/web_terminal.py b/core/web_terminal.py
index 7f7cbbd..964574d 100644
--- a/core/web_terminal.py
+++ b/core/web_terminal.py
@@ -393,10 +393,10 @@ class WebTerminal(MainTerminal):
                 "message": f"加载对话异常: {e}"
             }
     
-    def get_conversations_list(self, limit: int = 20, offset: int = 0, non_empty: bool = False) -> Dict:
+    def get_conversations_list(self, limit: int = 20, offset: int = 0, non_empty: bool = False, multi_agent_mode: Optional[bool] = None) -> Dict:
         """获取对话列表（Web版本）"""
         try:
-            result = self.context_manager.get_conversation_list(limit=limit, offset=offset, non_empty=non_empty)
+            result = self.context_manager.get_conversation_list(limit=limit, offset=offset, non_empty=non_empty, multi_agent_mode=multi_agent_mode)
             return {
                 "success": True,
                 "data": result
diff --git a/server/conversation.py b/server/conversation.py
index 4d97d5d..3c4db1c 100644
--- a/server/conversation.py
+++ b/server/conversation.py
@@ -525,6 +525,14 @@ def get_conversations(terminal: WebTerminal, workspace: UserWorkspace, username:
         limit = request.args.get('limit', 20, type=int)
         offset = request.args.get('offset', 0, type=int)
         non_empty = request.args.get('non_empty', '0') in ('1', 'true', 'True')
+        # multi_agent_mode: '1' 仅多智能体模式对话；'0' 仅常规对话；未传 None 不过滤
+        ma_param = request.args.get('multi_agent_mode', None)
+        if ma_param in ('1', 'true', 'True'):
+            multi_agent_filter: Optional[bool] = True
+        elif ma_param in ('0', 'false', 'False'):
+            multi_agent_filter = False
+        else:
+            multi_agent_filter = None
         target_workspace_id = request.args.get('workspace_id', '', type=str).strip()
 
         # 限制参数范围
@@ -541,7 +549,7 @@ def get_conversations(terminal: WebTerminal, workspace: UserWorkspace, username:
             except RuntimeError as exc:
                 return jsonify({"success": False, "error": str(exc)}), 503
 
-        result = terminal.get_conversations_list(limit=limit, offset=offset, non_empty=non_empty)
+        result = terminal.get_conversations_list(limit=limit, offset=offset, non_empty=non_empty, multi_agent_mode=multi_agent_filter)
         cm = getattr(getattr(terminal, "context_manager", None), "conversation_manager", None)
         write_host_workspace_debug(
             "sidebar-debug-api",
diff --git a/static/src/app/methods/conversation/load.ts b/static/src/app/methods/conversation/load.ts
index 50c172f..d0cbbd3 100644
--- a/static/src/app/methods/conversation/load.ts
+++ b/static/src/app/methods/conversation/load.ts
@@ -14,7 +14,9 @@ export const loadMethods = {
     const requestSeq = ++this.conversationListRequestSeq;
     this.conversationsLoading = true;
     try {
-      const response = await fetch(`/api/conversations?limit=${queryLimit}&offset=${queryOffset}`);
+      // 多智能体模式下只加载多智能体对话记录；常规模式排除多智能体对话
+      const maParam = this.multiAgentMode ? '&multi_agent_mode=1' : '&multi_agent_mode=0';
+      const response = await fetch(`/api/conversations?limit=${queryLimit}&offset=${queryOffset}${maParam}`);
       const data = await response.json();
 
       if (data.success) {
@@ -284,7 +286,8 @@ export const loadMethods = {
     if (group.loading || group.loadingMore) return;
     group.loading = true;
     try {
-      const response = await fetch(`/api/conversations?workspace_id=${encodeURIComponent(workspaceId)}&limit=${group.limit}&offset=${group.offset}`);
+      const maParam = this.multiAgentMode ? '&multi_agent_mode=1' : '&multi_agent_mode=0';
+      const response = await fetch(`/api/conversations?workspace_id=${encodeURIComponent(workspaceId)}&limit=${group.limit}&offset=${group.offset}${maParam}`);
       const data = await response.json();
       if (data.success) {
         const items = (data.data?.conversations || []).map((conv: any) => ({
diff --git a/utils/context_manager/conversation_mixin.py b/utils/context_manager/conversation_mixin.py
index 91e8cf2..d9b95f7 100644
--- a/utils/context_manager/conversation_mixin.py
+++ b/utils/context_manager/conversation_mixin.py
@@ -325,9 +325,9 @@ class ConversationMixin:
         except Exception as e:
             print(f"⌘ 自动保存异常: {e}")
 
-    def get_conversation_list(self, limit: int = 50, offset: int = 0, non_empty: bool = False) -> Dict:
+    def get_conversation_list(self, limit: int = 50, offset: int = 0, non_empty: bool = False, multi_agent_mode: Optional[bool] = None) -> Dict:
         """获取对话列表"""
-        return self.conversation_manager.get_conversation_list(limit=limit, offset=offset, non_empty=non_empty)
+        return self.conversation_manager.get_conversation_list(limit=limit, offset=offset, non_empty=non_empty, multi_agent_mode=multi_agent_mode)
 
     def delete_conversation_by_id(self, conversation_id: str) -> bool:
         """删除指定对话"""
diff --git a/utils/conversation_manager/list_search_mixin.py b/utils/conversation_manager/list_search_mixin.py
index f5c3fe1..9edea48 100644
--- a/utils/conversation_manager/list_search_mixin.py
+++ b/utils/conversation_manager/list_search_mixin.py
@@ -47,7 +47,7 @@ class ConversationMetadata:
 class ListSearchMixin:
     """ConversationManager list search mixin 能力 mixin。"""
 
-    def get_conversation_list(self, limit: int = 50, offset: int = 0, non_empty: bool = False) -> Dict:
+    def get_conversation_list(self, limit: int = 50, offset: int = 0, non_empty: bool = False, multi_agent_mode: Optional[bool] = None) -> Dict:
         """
         获取对话列表
 
@@ -55,14 +55,26 @@ class ListSearchMixin:
             limit: 限制数量
             offset: 偏移量
             non_empty: 仅返回有内容（total_messages > 0）的对话，分页基于过滤后的结果
+            multi_agent_mode: 若为 True 仅返回多智能体模式对话；若为 False 仅返回常规对话；None 不过滤
 
         Returns:
             Dict: 包含对话列表和统计信息
         """
         t0 = time.perf_counter()
         if perf_log:
-            perf_log("get_conversation_list enter", extra={"limit": limit, "offset": offset, "non_empty": non_empty})
+            perf_log("get_conversation_list enter", extra={"limit": limit, "offset": offset, "non_empty": non_empty, "multi_agent_mode": multi_agent_mode})
         try:
+            def _filter_by_multi_agent(items):
+                """按 multi_agent_mode 元数据过滤。"""
+                if multi_agent_mode is None:
+                    return items
+                result = []
+                for conv_id, meta in items:
+                    conv_ma = bool(meta.get("multi_agent_mode", False))
+                    if conv_ma == multi_agent_mode:
+                        result.append((conv_id, meta))
+                return result
+
             if non_empty:
                 # 过滤模式：全量加载后剔除空对话，再在过滤结果上分页，
                 # 保证 total / has_more 与"有内容对话"的真实数量一致。
@@ -76,6 +88,8 @@ class ListSearchMixin:
                     item for item in sorted_conversations
                     if (item[1].get("total_messages", 0) or 0) > 0
                 ]
+                if multi_agent_mode is not None:
+                    sorted_conversations = _filter_by_multi_agent(sorted_conversations)
                 total = len(sorted_conversations)
                 conversations = sorted_conversations[offset:offset + limit]
             else:
@@ -91,8 +105,11 @@ class ListSearchMixin:
                     reverse=True
                 )
 
+                if multi_agent_mode is not None:
+                    sorted_conversations = _filter_by_multi_agent(sorted_conversations)
+
                 # 分页
-                total = max(len(sorted_conversations), total_files)
+                total = max(len(sorted_conversations), total_files if multi_agent_mode is None else len(sorted_conversations))
                 conversations = sorted_conversations[offset:offset+limit]
 
             # 格式化结果
@@ -108,7 +125,8 @@ class ListSearchMixin:
                     "thinking_mode": metadata.get("thinking_mode", False),
                     "total_messages": metadata.get("total_messages", 0),
                     "total_tools": metadata.get("total_tools", 0),
-                    "status": metadata.get("status", "active")
+                    "status": metadata.get("status", "active"),
+                    "multi_agent_mode": bool(metadata.get("multi_agent_mode", False))
                 })
 
             elapsed_ms = (time.perf_counter() - t0) * 1000