diff --git a/packages/neon_framework/lib/src/utils/request_manager.dart b/packages/neon_framework/lib/src/utils/request_manager.dart
index df72b9c0a18..95006646c6d 100644
--- a/packages/neon_framework/lib/src/utils/request_manager.dart
+++ b/packages/neon_framework/lib/src/utils/request_manager.dart
@@ -11,6 +11,7 @@ import 'package:neon_framework/src/bloc/result.dart';
 import 'package:neon_framework/storage.dart';
 import 'package:neon_http_client/neon_http_client.dart';
 import 'package:nextcloud/utils.dart';
+import 'package:queue/queue.dart';
 import 'package:rxdart/rxdart.dart';
 import 'package:timezone/timezone.dart' as tz;
 
@@ -67,7 +68,12 @@ class RequestManager {
   @visibleForTesting
   http.Client? httpClient;
 
-  /// Executes a generic [http.Request].
+  /// This allows us to limit the amount of parallel requests to the server.
+  /// This is especially important when opening a gallery with a lot of images which require previews.
+  final _requestQueue = Queue(parallel: 10);
+
+  /// Executes a generic [http.Request] in a queued manner.
+  /// Only actual requests that are not cached will be executed against the server and queued.
   Future<void> wrap<T, R>({
     required Account account,
     required BehaviorSubject<Result<T>> subject,
@@ -79,11 +85,12 @@ class RequestManager {
     if (subject.isClosed) {
       return;
     }
+
     if (!subject.hasValue) {
       subject.add(Result.loading());
     }
 
-    var request = getRequest();
+    final request = getRequest();
 
     final cachedResponse = await _cache?.get(account, request);
     if (subject.isClosed) {
@@ -160,6 +167,29 @@ class RequestManager {
       subject.add(subject.value.asLoading());
     }
 
+    await _requestQueue.add(
+      () => _requestFromServer(
+        request: request,
+        account: account,
+        subject: subject,
+        getRequest: getRequest,
+        converter: converter,
+        unwrap: unwrap,
+        getCacheHeaders: getCacheHeaders,
+      ),
+    );
+  }
+
+  /// Performs [http.Request] against the server.
+  Future<void> _requestFromServer<T, R>({
+    required http.Request request,
+    required Account account,
+    required BehaviorSubject<Result<T>> subject,
+    required http.Request Function() getRequest,
+    required Converter<http.Response, R> converter,
+    required UnwrapCallback<T, R> unwrap,
+    AsyncValueGetter<Map<String, String>>? getCacheHeaders,
+  }) async {
     final client = httpClient ?? account.client;
 
     for (var i = 0; i < kMaxTries; i++) {
diff --git a/packages/neon_framework/packages/talk_app/test/room_bloc_test.dart b/packages/neon_framework/packages/talk_app/test/room_bloc_test.dart
index b6bbdb38f93..1f7bc6226d4 100644
--- a/packages/neon_framework/packages/talk_app/test/room_bloc_test.dart
+++ b/packages/neon_framework/packages/talk_app/test/room_bloc_test.dart
@@ -320,37 +320,66 @@ void main() {
 
   test('refresh', () async {
     expect(
-      roomBloc.room.transformResult((e) => e.token),
+      roomBloc.lastCommonRead,
+      emitsInOrder([0, 0]),
+    );
+
+    final initialRoom = expectLater(
+      roomBloc.room.transformResult((room) => room.token),
       emitsInOrder([
         Result.success('abcd').asLoading(),
         Result.success('abcd'),
         Result.success('abcd'),
+      ]),
+    );
+
+    final initialMessages = expectLater(
+      roomBloc.messages.transformResult(
+        (messages) => BuiltList<int>(messages.map((message) => message.id)),
+      ),
+      emitsInOrder([
+        Result<BuiltList<int>>.loading(),
+        Result.success(BuiltList<int>([2, 1, 0])),
+      ]),
+    );
+
+    await Future.wait([
+      initialRoom,
+      initialMessages,
+    ]);
+
+    final refreshedRoom = expectLater(
+      roomBloc.room
+          .skip(1) // we are skipping the initial refresh
+          .transformResult((room) => room.token),
+      emitsInOrder([
         Result.success('abcd').asLoading(),
         Result.success('abcd'),
         Result.success('abcd'),
       ]),
     );
 
-    expect(
-      roomBloc.messages.transformResult((e) => BuiltList<int>(e.map((m) => m.id))),
+    final refreshedMessages = expectLater(
+      roomBloc.messages
+          .skip(1) // we are skipping the initial refresh
+          .transformResult(
+            (messages) => BuiltList<int>(messages.map((message) => message.id)),
+          ),
       emitsInOrder([
-        Result<BuiltList<int>>.loading(),
-        Result.success(BuiltList<int>([2, 1, 0])),
         Result.success(BuiltList<int>([2, 1, 0])).asLoading(),
         Result.success(BuiltList<int>([2, 1, 0])),
       ]),
     );
 
-    expect(
-      roomBloc.lastCommonRead,
-      emitsInOrder([0, 0]),
-    );
-
-    // The delay is necessary to avoid a race condition with loading twice at the same time
-    await Future<void>.delayed(const Duration(milliseconds: 1));
     await roomBloc.refresh();
 
-    verify(() => talkBloc.updateRoom(any())).called(4);
+    await Future.wait([
+      refreshedRoom,
+      refreshedMessages,
+    ]);
+
+    // expecting 5 events: initial seed + 2 from constructor refresh + 2 from test refresh
+    verify(() => talkBloc.updateRoom(any())).called(5);
   });
 
   test('sendMessage', () async {
diff --git a/packages/neon_framework/pubspec.yaml b/packages/neon_framework/pubspec.yaml
index d2b931b89f2..9e7682a089d 100644
--- a/packages/neon_framework/pubspec.yaml
+++ b/packages/neon_framework/pubspec.yaml
@@ -57,6 +57,7 @@ dependencies:
   path_provider: ^2.1.0
   permission_handler: ^12.0.0
   provider: ^6.0.0
+  queue: ^3.0.0
   quick_actions: ^1.0.0
   rxdart: ^0.28.0
   scrollable_positioned_list: ^0.3.0
diff --git a/packages/neon_lints/pubspec.yaml b/packages/neon_lints/pubspec.yaml
index 536c4786aaf..e45b713d2e6 100644
--- a/packages/neon_lints/pubspec.yaml
+++ b/packages/neon_lints/pubspec.yaml
@@ -9,7 +9,7 @@ environment:
 
 dependencies:
   analyzer: ^7.3.0
-  analyzer_plugin: ^0.13.0
+  analyzer_plugin: ^0.13.1
   custom_lint_builder: ^0.7.0
 
 dev_dependencies: