diff --git a/common/arg.cpp b/common/arg.cpp
index c1151f51d..80f965cc7 100644
--- a/common/arg.cpp
+++ b/common/arg.cpp
@@ -1612,7 +1612,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
         [](common_params & params, const std::string & value) {
             params.antiprompt.emplace_back(value);
         }
-    ).set_examples({LLAMA_EXAMPLE_MAIN}));
+    ).set_examples({LLAMA_EXAMPLE_MAIN, LLAMA_EXAMPLE_SERVER}));
     add_opt(common_arg(
         {"-sp", "--special"},
         string_format("special tokens output enabled (default: %s)", params.special ? "true" : "false"),
diff --git a/tools/server/server.cpp b/tools/server/server.cpp
index 256a2928b..022b5d0b3 100644
--- a/tools/server/server.cpp
+++ b/tools/server/server.cpp
@@ -253,6 +253,7 @@ struct server_task {
         defaults.sampling    = params_base.sampling;
         defaults.speculative = params_base.speculative;
         defaults.n_keep      = params_base.n_keep;
+        defaults.antiprompt  = params_base.antiprompt;
 
         // enabling this will output extra debug information in the HTTP responses from the server
         params.verbose           = params_base.verbosity > 9;
@@ -490,6 +491,10 @@ struct server_task {
                     }
                 }
             }
+            // set reverse prompt from cli args if not set in the request
+            if (params.antiprompt.empty()) {
+                params.antiprompt = defaults.antiprompt;
+            }
         }
 
         {