diff --git a/README.md b/README.md
index e1606a5..9054dda 100644
--- a/README.md
+++ b/README.md
@@ -87,6 +87,12 @@ This has been tested with macOS, your mileage may vary on other operating system
    transcribe-me archive
    ```
 
+5. (Optional) You can also transcribe only the audio files that have not been transcribed yet:
+
+    ```bash
+    transcribe-me only
+    ```
+
 ### Docker
 
 You can also run the application using Docker:
diff --git a/transcribe_me/cli.py b/transcribe_me/cli.py
index 6304839..208683a 100644
--- a/transcribe_me/cli.py
+++ b/transcribe_me/cli.py
@@ -11,7 +11,7 @@ def parse_arguments():
     parser.add_argument(
         "command",
         nargs="?",
-        choices=["install", "archive"],
+        choices=["install", "archive", "only"],
         help="Install the configuration file or archive files.",
     )
     parser.add_argument(
@@ -40,6 +40,10 @@ def main():
         config_manager.archive_files(args.input, args.output)
         return
 
+    if args.command == "only":
+        transcription.process_audio_files(input_folder, output_folder, config)
+        return
+
     config = config_manager.load_config()
 
     input_folder = args.input