diff --git a/docs/admins_docs.md b/docs/admins_docs.md
index 356754b0..1b52b464 100644
--- a/docs/admins_docs.md
+++ b/docs/admins_docs.md
@@ -124,7 +124,9 @@ migrations_1     | Created admin user with password: gwg1clfkwf
 
 or if you have set the ADMIN_PASSWORD variable on docker-compose file you have used (example `docker-compose.yaml`), that variable will be set as the admin user's password
 
-`Note`: if you want to use the automatic transcriptions, run `docker-compose -f docker-compose.yaml -f docker-compose.full.yaml up` instead, as this is using a separate image.
+`Note`: if you want to use the automatic transcriptions, you have to do one of the following:
+* either use the docker-compose.full.yaml, so in this case run `docker-compose -f docker-compose.yaml -f docker-compose.full.yaml up`
+* or edit the docker-compose.yaml file and set the image for the celery_worker service as mediacms/mediacms:full instead of mediacms/mediacms:latest
 
 
 ### Update
@@ -1008,7 +1010,7 @@ When the whisper transcribe task is triggered for a media file, MediaCMS runs th
 
 ### Configuration
 
-Transcription functionality is available only for the Docker installation. To enable this feature, you must use the `docker-compose.full.yaml` file, as it contains an image with the necessary requirements.
+Transcription functionality is available only for the Docker installation. To enable this feature, you must either use the `docker-compose.full.yaml` file, as it contains an image with the necessary requirements, or you can also set that celery_worker service is usine mediacms:full image instead of mediacms:latest.
 
 By default, all users have the ability to send a request for a video to be transcribed, as well as transcribed and translated to English. If you wish to change this behavior, you can edit the `settings.py` file and set `USER_CAN_TRANSCRIBE_VIDEO=False`.
 
diff --git a/files/forms.py b/files/forms.py
index d2dc6aa4..510f64cf 100644
--- a/files/forms.py
+++ b/files/forms.py
@@ -1,6 +1,6 @@
 from crispy_forms.bootstrap import FormActions
 from crispy_forms.helper import FormHelper
-from crispy_forms.layout import Field, Layout, Submit
+from crispy_forms.layout import HTML, Field, Layout, Submit
 from django import forms
 from django.conf import settings
 
@@ -241,6 +241,8 @@ class WhisperSubtitlesForm(forms.ModelForm):
             self.fields['allow_whisper_transcribe_and_translate'].widget.attrs['readonly'] = True
             self.fields['allow_whisper_transcribe_and_translate'].widget.attrs['disabled'] = True
 
+        both_readonly = self.instance.allow_whisper_transcribe and self.instance.allow_whisper_transcribe_and_translate
+
         self.helper = FormHelper()
         self.helper.form_tag = True
         self.helper.form_class = 'post-form'
@@ -252,7 +254,13 @@ class WhisperSubtitlesForm(forms.ModelForm):
             CustomField('allow_whisper_transcribe_and_translate'),
         )
 
-        self.helper.layout.append(FormActions(Submit('submit_whisper', 'Submit', css_class='primaryAction')))
+        if not both_readonly:
+            self.helper.layout.append(FormActions(Submit('submit_whisper', 'Submit', css_class='primaryAction')))
+        else:
+            # Optional: Add a disabled button with explanatory text
+            self.helper.layout.append(
+                FormActions(Submit('submit_whisper', 'Submit', css_class='primaryAction', disabled=True), HTML('<small class="text-muted">Cannot submit - both options are already enabled</small>'))
+            )
 
     def clean_allow_whisper_transcribe(self):
         # Ensure the field value doesn't change if it was originally True
diff --git a/files/models/media.py b/files/models/media.py
index 82059743..1cbecfcc 100644
--- a/files/models/media.py
+++ b/files/models/media.py
@@ -235,6 +235,8 @@ class Media(models.Model):
         self.__original_media_file = self.media_file
         self.__original_thumbnail_time = self.thumbnail_time
         self.__original_uploaded_poster = self.uploaded_poster
+        self.__original_allow_whisper_transcribe = self.allow_whisper_transcribe
+        self.__original_allow_whisper_transcribe_and_translate = self.allow_whisper_transcribe_and_translate
 
     def save(self, *args, **kwargs):
         if not self.title:
@@ -275,6 +277,17 @@ class Media(models.Model):
             if self.thumbnail_time != self.__original_thumbnail_time:
                 self.__original_thumbnail_time = self.thumbnail_time
                 self.set_thumbnail(force=True)
+
+            transcription_changed = (
+                self.allow_whisper_transcribe != self.__original_allow_whisper_transcribe or self.allow_whisper_transcribe_and_translate != self.__original_allow_whisper_transcribe_and_translate
+            )
+
+            if transcription_changed and self.media_type == "video":
+                self.transcribe_function()
+
+            # Update the original values for next comparison
+            self.__original_allow_whisper_transcribe = self.allow_whisper_transcribe
+            self.__original_allow_whisper_transcribe_and_translate = self.allow_whisper_transcribe_and_translate
         else:
             # media is going to be created now
             # after media is saved, post_save signal will call media_init function
@@ -989,8 +1002,6 @@ def media_save(sender, instance, created, **kwargs):
             tag.update_tag_media()
 
     instance.update_search_vector()
-    if instance.media_type == "video":
-        instance.transcribe_function()
 
 
 @receiver(pre_delete, sender=Media)