diff --git a/tools/cli/cli.cpp b/tools/cli/cli.cpp
index 02ccb72598..ad421e6326 100644
--- a/tools/cli/cli.cpp
+++ b/tools/cli/cli.cpp
@@ -52,6 +52,7 @@ struct cli_context {
     json messages = json::array();
     std::vector<raw_buffer> input_files;
     task_params defaults;
+    bool verbose_prompt;
 
     // thread for showing "loading" animation
     std::atomic<bool> loading_show;
@@ -66,6 +67,8 @@ struct cli_context {
         defaults.stream = true; // make sure we always use streaming mode
         defaults.timings_per_token = true; // in order to get timings even when we cancel mid-way
         // defaults.return_progress = true; // TODO: show progress
+
+        verbose_prompt = params.verbose_prompt;
     }
 
     std::string generate_completion(result_timings & out_timings) {
@@ -91,6 +94,12 @@ struct cli_context {
             rd.post_task({std::move(task)});
         }
 
+        if (verbose_prompt) {
+            console::set_display(DISPLAY_TYPE_PROMPT);
+            console::log("%s\n\n", chat_params.prompt.c_str());
+            console::set_display(DISPLAY_TYPE_RESET);
+        }
+
         // wait for first result
         console::spinner::start();
         server_task_result_ptr result = rd.next(should_stop);