diff --git a/backend/open_webui/routers/ollama.py b/backend/open_webui/routers/ollama.py
index 7fc75996ae..b755acf80f 100644
--- a/backend/open_webui/routers/ollama.py
+++ b/backend/open_webui/routers/ollama.py
@@ -357,8 +357,10 @@ async def get_all_models(request: Request, user: UserModel | None = None):
 
     # Fan-out tag requests to every backend
     tasks = []
-    for idx, url in enumerate(await Config.get('ollama.base_urls', [])):
-        api_config = resolve_api_config((await Config.get('ollama.api_configs', {})), idx, url)
+    base_urls = await Config.get('ollama.base_urls', [])
+    api_configs = await Config.get('ollama.api_configs', {})
+    for idx, url in enumerate(base_urls):
+        api_config = resolve_api_config(api_configs, idx, url)
         if not api_config:
             tasks.append(send_get_request(f'{url}/api/tags', user=user))
         elif api_config.get('enable', True):
@@ -368,12 +370,16 @@ async def get_all_models(request: Request, user: UserModel | None = None):
 
     responses = await asyncio.gather(*tasks)
 
+    # Track which backends failed so we can skip them for /api/ps
+    failed_idxs: set[int] = set()
+
     # Post-process each response: apply prefix_id, tags, model filtering
     for idx, response in enumerate(responses):
         if not response:
+            failed_idxs.add(idx)
             continue
-        url = (await Config.get('ollama.base_urls', []))[idx]
-        api_config = resolve_api_config((await Config.get('ollama.api_configs', {})), idx, url)
+        url = base_urls[idx]
+        api_config = resolve_api_config(api_configs, idx, url)
 
         connection_type = api_config.get('connection_type', 'local')
         prefix_id = api_config.get('prefix_id')
@@ -395,7 +401,7 @@ async def get_all_models(request: Request, user: UserModel | None = None):
 
     # Annotate with expiry info from loaded-model state
     try:
-        loaded = await get_ollama_loaded_models(request, user=user)
+        loaded = await get_ollama_loaded_models(request, user=user, skip_idxs=failed_idxs)
         expires_map = {m['model']: m['expires_at'] for m in loaded['models'] if 'expires_at' in m}
         for m in models_dict['models']:
             if m['model'] in expires_map:
@@ -457,14 +463,20 @@ async def get_ollama_tags(
 async def get_ollama_loaded_models(
     request: Request,
     user=Depends(get_admin_user),
+    skip_idxs: set[int] | None = None,
 ) -> dict:
     """List models currently loaded in Ollama memory across all backends."""
     if not await Config.get('ollama.enable'):
         return {'models': []}
 
     tasks = []
-    for idx, url in enumerate(await Config.get('ollama.base_urls', [])):
-        api_config = resolve_api_config((await Config.get('ollama.api_configs', {})), idx, url)
+    base_urls = await Config.get('ollama.base_urls', [])
+    api_configs = await Config.get('ollama.api_configs', {})
+    for idx, url in enumerate(base_urls):
+        if skip_idxs and idx in skip_idxs:
+            tasks.append(asyncio.ensure_future(asyncio.sleep(0, None)))
+            continue
+        api_config = resolve_api_config(api_configs, idx, url)
         if not api_config:
             tasks.append(send_get_request(f'{url}/api/ps', user=user))
         elif api_config.get('enable', True):
@@ -477,7 +489,7 @@ async def get_ollama_loaded_models(
     for idx, response in enumerate(responses):
         if not response:
             continue
-        api_config = resolve_api_config((await Config.get('ollama.api_configs', {})), idx, (await Config.get('ollama.base_urls', []))[idx])
+        api_config = resolve_api_config(api_configs, idx, base_urls[idx])
         prefix_id = api_config.get('prefix_id')
         if prefix_id:
             for m in response.get('models', []):
diff --git a/src/lib/components/chat/MessageInput.svelte b/src/lib/components/chat/MessageInput.svelte
index 46cd0af3a7..4cdabb64f9 100644
--- a/src/lib/components/chat/MessageInput.svelte
+++ b/src/lib/components/chat/MessageInput.svelte
@@ -1545,7 +1545,7 @@
 														($settings?.promptAutocomplete ?? false)}
 													generateAutoCompletion={async (text) => {
 														if (selectedModelIds.length === 0 || !selectedModelIds.at(0)) {
-															toast.error($i18n.t('Please select a model first.'));
+															return null;
 														}
 
 														const res = await generateAutoCompletion(
diff --git a/src/routes/(app)/+layout.svelte b/src/routes/(app)/+layout.svelte
index f0e3a8dbbb..24ffc017dd 100644
--- a/src/routes/(app)/+layout.svelte
+++ b/src/routes/(app)/+layout.svelte
@@ -87,7 +87,7 @@
 		}
 	};
 
-	const setUserSettings = async (cb: () => Promise<void>) => {
+	const setUserSettings = async (cb?: () => Promise<void>) => {
 		let userSettings = await getUserSettings(localStorage.token).catch((error) => {
 			console.error(error);
 			return null;
@@ -209,14 +209,15 @@
 			checkLocalDBChats(),
 			setBanners().catch((e) => console.error('Failed to load banners:', e)),
 			setTools().catch((e) => console.error('Failed to load tools:', e)),
-			setUserSettings(async () => {
-				await Promise.all([
-					setModels().catch((e) => console.error('Failed to load models:', e)),
-					setToolServers().catch((e) => console.error('Failed to load tool servers:', e))
-				]);
-			}).catch((e) => console.error('Failed to load user settings:', e))
+			setUserSettings().catch((e) => console.error('Failed to load user settings:', e))
 		]);
 
+		// Load models and tool servers in the background — don't block page render.
+		// These contact external services (Ollama, OpenAI, tool servers, terminal
+		// servers) that may be slow or unreachable.
+		setModels().catch((e) => console.error('Failed to load models:', e));
+		setToolServers().catch((e) => console.error('Failed to load tool servers:', e));
+
 		// Helper function to check if the pressed keys match the shortcut definition
 		const isShortcutMatch = (event: KeyboardEvent, shortcut): boolean => {
 			const keys = shortcut?.keys || [];