feat: implement LLM backend management support with new CJS server and frontend components

app/frontend/src/App.jsx

@@ -122,7 +122,7 @@ function App() {
       } catch (_) {}
     }
     return {
-      contextSize: 4096,
+      contextSize: 0,
       temperature: 0.7,
       systemPrompt: "You are a helpful local AI assistant.",
       threads: 4

app/frontend/src/components/Settings.jsx

@@ -512,8 +512,9 @@ function Settings({
                   className="m3-input"
                   style={{ marginTop: "6px", height: "40px", cursor: "pointer" }}
                 >
+                  <option value="0">Auto-detect (Recommended)</option>
                   <option value="2048">2,048 tokens (Fast/Low Memory)</option>
-                  <option value="4096">4,096 tokens (Balanced / Recommended)</option>
+                  <option value="4096">4,096 tokens (Balanced)</option>
                   <option value="8192">8,192 tokens (More history)</option>
                   <option value="16384">16,384 tokens (Deep context)</option>
                 </select>