Sync from internal repo (YYYY/MM/DD)

package.json

@@ -1,6 +1,6 @@
 {
   "name": "assemblyai",
-  "version": "4.24.0",
+  "version": "4.25.0",
   "description": "The AssemblyAI JavaScript SDK provides an easy-to-use interface for interacting with the AssemblyAI API, which supports async and real-time transcription, as well as the latest LeMUR models.",
   "engines": {
     "node": ">=18"

src/services/streaming/service.ts

@@ -165,6 +165,14 @@ export class StreamingTranscriber {
       );
     }
 
+    if (this.params.speakerLabels !== undefined) {
+      searchParams.set("speaker_labels", this.params.speakerLabels.toString());
+    }
+
+    if (this.params.maxSpeakers !== undefined) {
+      searchParams.set("max_speakers", this.params.maxSpeakers.toString());
+    }
+
     url.search = searchParams.toString();
 
     return url;

src/types/streaming/index.ts

@@ -22,6 +22,8 @@ export type StreamingTranscriberParams = {
   speechModel?: StreamingSpeechModel;
   languageDetection?: boolean;
   inactivityTimeout?: number;
+  speakerLabels?: boolean;
+  maxSpeakers?: number;
 };
 
 export type StreamingEvents = "open" | "close" | "turn" | "error";
@@ -37,6 +39,7 @@ export type StreamingSpeechModel =
   | "universal-streaming-english"
   | "universal-streaming-multilingual"
   | "u3-rt-pro"
+  | "whisper-rt"
   | "u3-pro";
 
 export type StreamingTokenParams = {
@@ -66,6 +69,7 @@ export type TurnEvent = {
   words: StreamingWord[];
   language_code?: string;
   language_confidence?: number;
+  speaker_label?: string;
 };
 
 export type StreamingWord = {

tests/unit/streaming.test.ts

@@ -59,4 +59,76 @@ describe("streaming", () => {
   }
 
   it("noop", async () => {});
+
+  it("should include speaker_labels in connection URL", async () => {
+    await cleanup();
+    WS.clean();
+
+    const wsUrl = `${websocketBaseUrl}?token=123&sample_rate=16000&speaker_labels=true`;
+    server = new WS(wsUrl);
+    rt = new StreamingTranscriber({
+      websocketBaseUrl,
+      token: "123",
+      sampleRate: 16_000,
+      speakerLabels: true,
+    });
+    onOpen = jest.fn();
+    rt.on("open", onOpen);
+    await connect(rt, server);
+  });
+
+  it("should include speaker_labels and max_speakers in connection URL", async () => {
+    await cleanup();
+    WS.clean();
+
+    const wsUrl = `${websocketBaseUrl}?token=123&sample_rate=16000&speaker_labels=true&max_speakers=4`;
+    server = new WS(wsUrl);
+    rt = new StreamingTranscriber({
+      websocketBaseUrl,
+      token: "123",
+      sampleRate: 16_000,
+      speakerLabels: true,
+      maxSpeakers: 4,
+    });
+    onOpen = jest.fn();
+    rt.on("open", onOpen);
+    await connect(rt, server);
+  });
+
+  it("should include whisper-rt speech model in connection URL", async () => {
+    await cleanup();
+    WS.clean();
+
+    const wsUrl = `${websocketBaseUrl}?token=123&sample_rate=16000&speech_model=whisper-rt`;
+    server = new WS(wsUrl);
+    rt = new StreamingTranscriber({
+      websocketBaseUrl,
+      token: "123",
+      sampleRate: 16_000,
+      speechModel: "whisper-rt",
+    });
+    onOpen = jest.fn();
+    rt.on("open", onOpen);
+    await connect(rt, server);
+  });
+
+  it("should parse speaker_label from turn event", async () => {
+    const turnPromise = new Promise<{ speaker_label?: string }>((resolve) => {
+      rt.on("turn", (event) => resolve(event));
+    });
+    server.send(
+      JSON.stringify({
+        type: "Turn",
+        turn_order: 1,
+        turn_is_formatted: true,
+        end_of_turn: true,
+        transcript: "hello",
+        end_of_turn_confidence: 0.9,
+        words: [],
+        speaker_label: "A",
+      }),
+    );
+    const turn = await turnPromise;
+    expect(turn.speaker_label).toBe("A");
+  });
 });