diff --git a/zerver/lib/message.py b/zerver/lib/message.py
index e062642436..60d7d1367c 100644
--- a/zerver/lib/message.py
+++ b/zerver/lib/message.py
@@ -1010,6 +1010,9 @@ def extract_unread_data_from_um_rows(
         if topic_mute_checker(recipient_id, topic):
             return True
 
+        # Messages sent by muted users are never unread, so we don't
+        # need any logic related to muted users here.
+
         return False
 
     huddle_cache: Dict[int, str] = {}
diff --git a/zerver/views/message_fetch.py b/zerver/views/message_fetch.py
index f50f446591..566c175588 100644
--- a/zerver/views/message_fetch.py
+++ b/zerver/views/message_fetch.py
@@ -736,6 +736,9 @@ def exclude_muting_conditions(
     except Stream.DoesNotExist:
         pass
 
+    # Stream-level muting only applies when looking at views that
+    # include multiple streams, since we do want users to be able to
+    # browser messages within a muted stream.
     if stream_id is None:
         rows = Subscription.objects.filter(
             user_profile=user_profile,
@@ -751,6 +754,15 @@ def exclude_muting_conditions(
 
     conditions = exclude_topic_mutes(conditions, user_profile, stream_id)
 
+    # Muted user logic for hiding messages is implemented entirely
+    # client-side. This is by design, as it allows UI to hint that
+    # muted messages exist where their absence might make conversation
+    # difficult to understand. As a result, we do not need to consider
+    # muted users in this server-side logic for returning messages to
+    # clients. (We could in theory exclude PMs from muted users, but
+    # they're likely to be sufficiently rare to not be worth extra
+    # logic/testing here).
+
     return conditions