Greet users when voice sessions start

main.js

@@ -25,6 +25,9 @@ const DEFAULT_VOICE = "Aiden";
 const DEFAULT_INSTRUCTIONS =
   "You are a friendly voice assistant. " +
   "Keep replies short, warm, and spoken. Avoid long monologues.";
+const STARTUP_GREETING_PROMPT =
+  "Start the conversation now with a brief, spontaneous greeting in character. " +
+  "Keep it to one sentence, invite the user in naturally, and vary the wording each time.";
 
 // Appended to the user's instructions whenever at least one tool is enabled.
 // Stops the model from announcing capabilities ("Yes, I can search") and then
@@ -1183,6 +1186,7 @@ async function doStart(audioContext = null) {
     ...target,
     voice: settings.voice,
     instructions: effectiveInstructions(),
+    startupGreeting: STARTUP_GREETING_PROMPT,
     acquireMic: acquireMicStream,
     tools: activeToolDefs(),
     noiseGate: gateParams(settings.noiseGate),

ws/s2s-ws-client.js

@@ -51,6 +51,8 @@
  *   session POST and dials it directly — no load balancer in between.
  * @property {string} voice
  * @property {string} instructions
+ * @property {string} [startupGreeting] Hidden user prompt that asks the model
+ *   to greet once after the initial session configuration is acknowledged.
  * @property {MediaStream} [micStream] Live mic stream. Provide this OR `acquireMic`.
  * @property {() => Promise<MediaStream>} [acquireMic] Lazily obtain the mic stream,
  *   called only once a session is actually granted (after any queue wait). Lets the
@@ -198,6 +200,8 @@ export class S2sWsRealtimeClient extends EventTarget {
     /** @type {Promise<void> | null} */
     this._readyPromise = null;
     this._sessionConfigured = false;
+    this._startupGreeting = options.startupGreeting?.trim() ?? "";
+    this._startupGreetingSent = false;
     this._debug = (() => { try { return localStorage.getItem("s2s.debug") === "1"; } catch { return false; } })();
   }
 
@@ -629,7 +633,9 @@ export class S2sWsRealtimeClient extends EventTarget {
         break;
 
       case "session.updated":
-        // Acknowledged by server, nothing to do.
+        // Wait for this acknowledgement so the real system instructions are in
+        // place before the greeting request warms their prompt prefix.
+        this._sendStartupGreeting();
         break;
 
       case "input_audio_buffer.speech_started":
@@ -965,6 +971,26 @@ export class S2sWsRealtimeClient extends EventTarget {
     });
   }
 
+  /**
+   * Ask the model to open the conversation exactly once. The synthetic user
+   * prompt stays in conversation history, so the greeting warms the actual
+   * system-prompt prefix that the first spoken turn will reuse.
+   */
+  _sendStartupGreeting() {
+    if (!this._startupGreeting || this._startupGreetingSent) return;
+    this._startupGreetingSent = true;
+    this._send({
+      type: "conversation.item.create",
+      item: {
+        type: "message",
+        role: "user",
+        content: [{ type: "input_text", text: this._startupGreeting }],
+      },
+    });
+    this.requestResponse();
+    if (this._debug) console.debug("[ws] startup greeting queued");
+  }
+
   /**
    * Ask the model to generate a response now (after feeding tool results).
    * Serialized: if a response is already in flight we queue this request and