SDKsHarmonyOSEnterprise

Transcribe audio to text

Convert Base64-encoded audio data into recognized text.

Do not pass local file paths, raw ArrayBuffer instances, or data URLs prefixed with data:*;base64,.

Code example

import sdk, { SpeechToTextReq, SpeechToTextResp } from '@openimsdk/imsdk';

const request: SpeechToTextReq = {
  filename: 'voice.m4a',
  data: audioBase64,
};
const response: SpeechToTextResp = await sdk.speechToText(request);
showTranscript(response.text);

Parameters

Parameter	Type	Required	Description
`filename`	`string`	Yes	Audio filename including the valid file extension.
`data`	`string`	Yes	Pure Base64-encoded audio payload string.

Return value

Returns a Promise resolving to SpeechToTextResp:

Property	Type	Description
`text`	`string`	Recognized text content extracted from the audio.

filename must include an extension corresponding to the real audio format, and data must contain only the raw Base64 string. Before calling this method, validate format, sampling rate, duration, and file size limits against Check speech-to-text capabilities. Transcription results do not modify the original voice message record or trigger message events.

Was this page helpful?

Code example

Parameters

Return value

State changes after the call