SDK Reference

This reference guide provides detailed information about all classes, methods, interfaces, and options available in the Klarisent STT SDK.

KlarisentSTT Class

The main class for interacting with the Klarisent Speech-to-Text service.

Constructor

constructor({
  api_key,
  connectionId,
  debug = false,
  pauseDuration = 0.6,
}: IKlarisentSTTConstructor)

Parameters

Parameter	Type	Description
api_key	string	Required. Your Klarisent API key
connectionId	string	Optional. Custom identifier for the connection. If not provided, a UUID will be generated
debug	boolean	Optional. Enable debug logging. Default: false
pauseDuration	number	Optional. Pause duration for Voice Activity Detection in seconds. Minimum: 0.5. Default: 0.6

Methods

establishConnection

Establishes a WebSocket connection to the Klarisent server.

establishConnection({ replaceableObjects }: { replaceableObjects: [] }): Promise<{ status: number, message: string }>

Parameters:

• replaceableObjects: An array of objects (reserved for future use)

Returns:

• Promise resolving to a connection status object

Example:

await stt.establishConnection({ replaceableObjects: [] });

sendBase64Audio

Sends base64-encoded audio for transcription through the REST API.

sendBase64Audio(payload: ISendBase64AudioPayload): Promise<ISendAudioStreamResponse>

Parameters:

• payload: Object containing audio data and configuration options

Returns:

• Promise resolving to a transcription response object

Example:

const result = await stt.sendBase64Audio({
  audio: base64String,
  fileName: 'recording.flac',
  language: 'en',
  trigger: [TRIGGER.QuestionDetection]
});

sendAudioStream

Streams audio for real-time transcription through WebSockets.

sendAudioStream(audioStream: any): void

Parameters:

• audioStream: A readable stream of audio data

Returns:

• void

Example:

stt.sendAudioStream(microphoneStream);

stop

Terminates the connection.

stop(): void

Returns:

• void

Example:

stt.stop();

onTranscription

Registers a callback function for transcription events.

onTranscription(callback: (data: ITranscriptionCallBackPayload) => void): void

Parameters:

• callback: Function to be called when transcription is received

Example:

stt.onTranscription((data) => {
  console.log('Transcription:', data.transcript);
});

onQuestion

Registers a callback function for question detection events.

onQuestion(callback: (data: IQuestionCallBackPayload) => void): void

Parameters:

• callback: Function to be called when questions are detected

Example:

stt.onQuestion((data) => {
  console.log('Questions:', data.questions);
});

onError

Registers a callback function for error events.

onError(callback: (error: any) => void): void

Parameters:

• callback: Function to be called when an error occurs

Example:

stt.onError((error) => {
  console.error('Error:', error);
});

onStreamEnd

Registers a callback function for stream end events.

onStreamEnd(callback: () => void): void

Parameters:

• callback: Function to be called when the audio stream ends

Example:

stt.onStreamEnd(() => {
  console.log('Stream ended');
});

setDebugMode

Enables or disables debug logging.

setDebugMode(debug: boolean): void

Parameters:

• debug: Boolean indicating whether to enable debug logging

Example:

stt.setDebugMode(true);

Properties

isConnected

Indicates if the WebSocket connection is currently established.

isConnected: boolean

Example:

console.log('Connection status:', stt.isConnected);

Interfaces

IKlarisentSTTConstructor

Parameters for initializing the KlarisentSTT class.

interface IKlarisentSTTConstructor {
  api_key: string;
  connectionId?: string;
  debug?: boolean;
  pauseDuration?: number;
}

ISendBase64AudioPayload

Parameters for the sendBase64Audio method.

interface ISendBase64AudioPayload {
  audio: string;
  fileName?: string;
  language: "en" | "hi";
  trigger?: TRIGGER[];
}

ITranscriptionCallBackPayload

Data received in transcription events.

interface ITranscriptionCallBackPayload {
  transcript: string;
  questions: string[];
}

IQuestionCallBackPayload

Data received in question detection events.

interface IQuestionCallBackPayload {
  transcript: string;
  questions: string;
}

ISendAudioStreamResponse

Response from the sendBase64Audio method.

interface ISendAudioStreamResponse {
  status: number;
  message: string;
  data: ISendAudioStreamResponseData;
}

ISendAudioStreamResponseData

Data included in the API response.

interface ISendAudioStreamResponseData {
  audioSizeKB: number;
  duration: number;
  transcription: string;
  transcriptionTime: number;
  trigger?: {
    [TRIGGER.QuestionDetection]?: string[];
    [TRIGGER.SentimentAnalysis]?: {
      sentiment: string;
      text: string;
    };
    [TRIGGER.SentimentAnalysisAdvanced]?: {
      sentiment: string;
      text: string;
    };
  };
}

Enums

TRIGGER

Enumeration of available analysis triggers.

enum TRIGGER {
  QuestionDetection = "6746edc31c1965a1ee9bc2d4",
  SentimentAnalysis = "674882924ed3a11958beaab7",
  SentimentAnalysisAdvanced = "67605b1a4fcedccfa3e2d494"
}

Constants

BASE_URL

The base URL for API requests.

const BASE_URL = "https://api.klarisent.com";

Error Codes

Status Code	Description
200	Success
400	Bad Request - Invalid parameters
401	Unauthorized - Invalid API key
403	Forbidden - Account restrictions
413	Payload Too Large - Audio file too big
429	Too Many Requests - Rate limit exceeded
500	Internal Server Error

Best Practices

Audio Formats

For optimal recognition:

• Sample Rate: 16000 Hz

• Channels: 1 (mono)

• Format: FLAC or WAV

• Bit Depth: 16-bit

Connection Management

• Always check isConnected before sending audio streams

• Handle reconnection logic for