Sync from internal repo (2025/10/16)

package.json

@@ -1,6 +1,6 @@
 {
   "name": "assemblyai",
-  "version": "4.17.0",
+  "version": "4.18.0",
   "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/types/openapi.generated.ts

@@ -1037,7 +1037,7 @@
   /**
    * Only get throttled transcripts, overrides the status filter
    * @defaultValue false
    * @deprecated
    */
   throttled_only?: boolean;
 };
@@ -1481,6 +1481,59 @@
   medium: number;
 };
 
+/**
+ * Speaker identification type for speech understanding
+ */
+export type SpeakerType = "role" | "name";
+
+/**
+ * Speaker identification configuration for speech understanding
+ */
+export type SpeakerIdentificationRequest = {
+  /**
+   * The type of speaker identification to perform
+   */
+  speaker_type: SpeakerType;
+  /**
+   * Known speaker values (required if speaker_type is 'role')
+   */
+  known_values?: string[];
+};
+
+/**
+ * Speech understanding request configuration
+ */
+export type SpeechUnderstandingRequest = {
+  /**
+   * Speaker identification configuration
+   */
+  speaker_identification?: SpeakerIdentificationRequest;
+};
+
+/**
+ * Status of a speech understanding feature
+ */
+export type SpeechUnderstandingFeatureStatus = {
+  /**
+   * Status of the feature (e.g., 'success')
+   */
+  status: string;
+};
+
+/**
+ * Speech understanding response containing feature statuses and the original request
+ */
+export type SpeechUnderstandingResponse = {
+  /**
+   * Status of speaker identification feature
+   */
+  speaker_identification?: SpeechUnderstandingFeatureStatus;
+  /**
+   * The original speech understanding request
+   */
+  request?: SpeechUnderstandingRequest;
+};
+
 /**
  * Advanced options for controlling speaker diarization parameters
  */
@@ -2622,6 +2675,14 @@
    * @defaultValue "null
    */
   speech_model: SpeechModel | null;
+  /**
+   * The list of speech models to use for the transcription in priority order.
+   */
+  speech_models?: string[] | null;
+  /**
+   * The actual speech model that was used for the transcription.
+   */
+  speech_model_used?: string | null;
   /**
    * Defaults to null. Reject audio files that contain less than this fraction of speech.
    * Valid values are in the range [0", 1] inclusive.
@@ -3066,6 +3127,12 @@
    * The list of key terms used to generate the transcript with the Slam-1 speech model. Can't be used together with `prompt`.
    */
   keyterms_prompt?: string[];
+  /**
+   * Speech understanding configuration/response for speaker identification
+   */
+  speech_understanding?:
+    | SpeechUnderstandingRequest
+    | SpeechUnderstandingResponse;
   /**
    * The language of your audio file. Possible values are found in {@link https://www.assemblyai.com/docs/concepts/supported-languages | Supported Languages }.
    * The default value is 'en_us'.
@@ -3152,6 +3219,10 @@
    * @defaultValue best
    */
   speech_model?: SpeechModel | null;
+  /**
+   * The list of speech models to use for the transcription in priority order.
+   */
+  speech_models?: string[] | null;
   /**
    * Reject audio files that contain less than this fraction of speech.
    * Valid values are in the range [0", 1] inclusive.

tests/unit/transcript.test.ts

@@ -468,4 +468,39 @@ describe("transcript", () => {
     expect(searchResponse.total_count).toBe(1);
     expect(searchResponse.matches).toBeInstanceOf(Array);
   });
+
+  it("should handle transcript response with speech_model_used field", async () => {
+    const transcriptWithSpeechModelUsed = {
+      id: transcriptId,
+      status: "completed",
+      speech_model_used: "best",
+    };
+    fetchMock.doMockOnceIf(
+      requestMatches({ url: `/v2/transcript/${transcriptId}`, method: "GET" }),
+      JSON.stringify(transcriptWithSpeechModelUsed),
+    );
+    const transcript = await assembly.transcripts.get(transcriptId);
+
+    expect(transcript.id).toBe(transcriptId);
+    expect(transcript.speech_model_used).toBe("best");
+  });
+
+  it("should handle transcript response without speech_model_used field", async () => {
+    // This test verifies that the SDK gracefully handles responses
+    // where speech_model_used is not present, as the field has not yet
+    // been added to the API for all users.
+    const transcriptWithoutSpeechModelUsed = {
+      id: transcriptId,
+      status: "completed",
+      // speech_model_used intentionally omitted
+    };
+    fetchMock.doMockOnceIf(
+      requestMatches({ url: `/v2/transcript/${transcriptId}`, method: "GET" }),
+      JSON.stringify(transcriptWithoutSpeechModelUsed),
+    );
+    const transcript = await assembly.transcripts.get(transcriptId);
+
+    expect(transcript.id).toBe(transcriptId);
+    expect(transcript.speech_model_used).toBeUndefined();
+  });
 });