Worktual Voice STT SDK

Upload audio and receive transcript text. Node accepts a filesystem path. The browser build accepts a File or Blob.

@worktual_limited/stt

Install

npm i @worktual_limited/stt

Node

import { SpeechClient } from "@worktual_limited/stt";

const client = new SpeechClient({
  accessToken: "identity-jwt",
  superDomainId: "YOUR_SUPER_DOMAIN",
  subDomainId: "1000",
  baseUrl: "https://unified-api.worktual.com",
});

const text = await client.transcribe("calls/call1.wav");

Browser

import { SpeechClient } from "@worktual_limited/stt";

const client = new SpeechClient({
  accessToken: "identity-jwt",
  superDomainId: "YOUR_SUPER_DOMAIN",
  subDomainId: "1000",
  baseUrl: "https://unified-api.worktual.com",
});

const text = await client.transcribe(file);

Bundlers select the browser build, which takes a File or Blob. transcribe posts multipart form data to POST {baseUrl}/stt. The file field is audio. The response text field is the transcript. subDomainId is the numeric subdomain id from My Account. A failed response throws SpeechError with status and the response body.

Do not ship STT access credentials in the browser for production apps — prefer a server/BFF.

Test

npm test

In this document

Worktual Voice STT SDK

Install

Node

Browser

Test

↑ Back to top