Sync from GitHub 2026-04-30T05:47:36Z

index.html

@@ -2176,8 +2176,8 @@
         <div class="welcome" id="welcomeMsg">
           <div class="welcome-icon">&#129504;</div>
           <h2>Private AI, right in your browser</h2>
-          <p>Gemma runs entirely on your device. Nothing is sent to any server.</p>
-          <p style="font-size:0.78rem;color:var(--gray-400);margin-top:4px">Switch to a Gemma 4 model for tool calling, memory, and multimodal input.<br>Web search requires an API key (Tavily, Brave) or a self-hosted SearXNG instance.</p>
+          <p>Models run entirely on your device via WebGPU. Nothing is sent to any server.</p>
+          <p style="font-size:0.78rem;color:var(--gray-400);margin-top:4px">Pick a Ternary Bonsai or Qwen3 model for tool calling and memory; pick a Gemma 4 model for multimodal (image + audio) input.<br>Web search requires an API key (Tavily, Brave) or a self-hosted SearXNG instance.</p>
         </div>
       </div>