feat(nodesdk): Typed support for SU understanding transcript config api

CHANGELOG.md

@@ -1,5 +1,9 @@
 # Changelog
 
+## [4.41.7]
+
+- Add `summarization` to the speech understanding feature requests and responses (`SummarizationRequest`, `SummarizationResponse`): `summary_type`, `headline_length`, `approximate_chapters` and `effort`. The feature was already accepted by the API but missing from the SDK types, so setting it failed to type check
+
 ## [4.41.6]
 
 - Add `rewrite_mode` to `DictationConfig` — the model tier that runs the dictation rewrite: `"fast"` (the server default) or `"smart"`, which follows detailed formatting instructions more reliably at higher latency. Unset, it is not sent. The `DictationRewriteMode` type is exported from the package root

package.json

@@ -1,6 +1,6 @@
 {
   "name": "assemblyai",
-  "version": "4.41.6",
+  "version": "4.41.7",
   "description": "The AssemblyAI JavaScript SDK provides an easy-to-use interface for interacting with the AssemblyAI API, which supports async and real-time transcription.",
   "engines": {
     "node": ">=18"

src/types/openapi.generated.ts

@@ -1120,6 +1120,37 @@ export type CustomFormattingRequest = {
   effort?: SpeechUnderstandingEffort;
 };
 
+/**
+ * Summarization type for speech understanding
+ */
+export type SummarizationType = "bullets" | "paragraph";
+
+/**
+ * Summarization configuration for speech understanding
+ */
+export type SummarizationRequest = {
+  /**
+   * The format of the summary
+   */
+  summary_type: SummarizationType;
+  /**
+   * Target word count for each chapter headline (5-15)
+   *
+   * @defaultValue 10
+   */
+  headline_length?: number;
+  /**
+   * Override: Approximate number of chapters to split the transcript into
+   */
+  approximate_chapters?: number;
+  /**
+   * How much effort to spend on this task
+   *
+   * @defaultValue "low"
+   */
+  effort?: SpeechUnderstandingEffort;
+};
+
 /**
  * Speech understanding feature requests grouped together
  */
@@ -1136,6 +1167,10 @@ export type SpeechUnderstandingFeatureRequests = {
    * Custom formatting configuration
    */
   custom_formatting?: CustomFormattingRequest;
+  /**
+   * Summarization configuration
+   */
+  summarization?: SummarizationRequest;
 };
 
 /**
@@ -1194,6 +1229,58 @@ export type CustomFormattingResponse = {
   status: string;
 };
 
+/**
+ * A summarized section of the transcript
+ */
+export type SummarizationSection = {
+  /**
+   * Start of the section, in milliseconds
+   */
+  start: number;
+  /**
+   * End of the section, in milliseconds
+   */
+  end: number;
+  /**
+   * Section summary, when summary_type is 'paragraph'
+   */
+  text?: string;
+  /**
+   * Section summary, when summary_type is 'bullets'
+   */
+  bullets?: string[];
+  /**
+   * Short headline describing the section
+   */
+  headline?: string;
+};
+
+/**
+ * Summarization response containing the chapter summaries
+ */
+export type SummarizationResponse = {
+  /**
+   * Status of the summarization feature
+   */
+  status: string;
+  /**
+   * Per-section summaries
+   */
+  summary?: SummarizationSection[];
+  /**
+   * Single summary covering the whole transcript
+   */
+  block_summary?: string;
+  /**
+   * The summary type that was used
+   */
+  summary_type?: SummarizationType;
+  /**
+   * The effort level that was used
+   */
+  effort?: SpeechUnderstandingEffort;
+};
+
 /**
  * Speech understanding feature responses grouped together
  */
@@ -1210,6 +1297,10 @@ export type SpeechUnderstandingFeatureResponses = {
    * Custom formatting results
    */
   custom_formatting?: CustomFormattingResponse;
+  /**
+   * Summarization results
+   */
+  summarization?: SummarizationResponse;
 };
 
 /**

tests/unit/speech-understanding-summarization.test.ts

@@ -0,0 +1,52 @@
+import fetchMock from "jest-fetch-mock";
+import { SummarizationRequest, SummarizationResponse } from "../../src";
+import { createClient, requestMatches } from "./utils";
+
+fetchMock.enableMocks();
+
+const assembly = createClient();
+const transcriptId = "transcript_123";
+const remoteAudioURL = "https://assembly.ai/espn.m4a";
+
+beforeEach(() => {
+  jest.clearAllMocks();
+  fetchMock.resetMocks();
+  fetchMock.doMock();
+});
+
+describe("speech understanding summarization", () => {
+  it("should submit every summarization option", async () => {
+    const summarization: SummarizationRequest = {
+      summary_type: "bullets",
+      headline_length: 8,
+      approximate_chapters: 4,
+      effort: "medium",
+    };
+    fetchMock.doMockOnceIf(
+      requestMatches({ url: "/v2/transcript", method: "POST" }),
+      JSON.stringify({ id: transcriptId, status: "queued" }),
+    );
+
+    await assembly.transcripts.submit({
+      audio_url: remoteAudioURL,
+      speech_understanding: { request: { summarization } },
+    });
+
+    const body = JSON.parse(fetchMock.mock.calls[0][1]?.body as string);
+    expect(body.speech_understanding.request.summarization).toEqual(
+      summarization,
+    );
+  });
+
+  it("should type the summarization response", () => {
+    const response: SummarizationResponse = {
+      status: "success",
+      summary_type: "paragraph",
+      effort: "low",
+      block_summary: "A summary.",
+      summary: [{ start: 0, end: 1000, text: "A summary.", headline: "Intro" }],
+    };
+
+    expect(response.summary?.[0].headline).toBe("Intro");
+  });
+});