@talkscriber-npm/ts-client NPM

About ts-client

ts-client is the official TypeScript client for Talkscriber, a state-of-the-Art Speech-to-Text (STT) platform tailored for conversational AI enterprises. It provides exceptional transcription services with a strong emphasis on privacy and security to enhance customer communication while protecting sensitive information.

Key Features of Talkscriber

A Word Error Rate (WER) of less than 4%
Very low latency (under 150 ms)
Support for 50+ languages

Installation and Getting Started

Follow these steps to install and use the ts-client for Talkscriber:

Install the package:
```
npm install @talkscriber-npm/ts-client
```

In your project, create a new file (e.g., transcribe.ts) and add the following code:

import { TalkscriberTranscriptionService } from '@talkscriber-npm/ts-client';

async function main() {
  const talkscriber = new TalkscriberTranscriptionService({
    apiKey: '<YOUR_API_KEY>',
    language: 'en', // Specify the language here (e.g., 'en' for English)
    onTranscription: (text: string) => {
      console.log('Transcription:', text);
    },
    onUtterance: (text: string) => {
      console.log('Utterance:', text);
    }
  });

  try {
    await talkscriber.connect();
    console.log('Connected to Talkscriber service');

    // Your audio processing code here
    // For example:
    // const audioData = new Float32Array(/* your audio data */);
    // const sampleRate = 44100; // or your actual sample rate
    // talkscriber.send(audioData, sampleRate);

  } catch (error) {
    console.error('Error:', error instanceof Error ? error.message : String(error));
  } finally {
    talkscriber.close();
  }
}

main().catch(console.error);

Replace <YOUR_API_KEY> with your actual Talkscriber API key.
Install the necessary TypeScript dependencies if you haven't already:
```
npm install -D typescript ts-node @types/node
```
Compile and run your TypeScript code:
```
npx ts-node transcribe.ts
```

This will initialize the Talkscriber client and connect to the service. You'll need to add your own audio processing logic to send audio data to the talkscriber.send() method.

For complete examples of audio file processing, refer to the examples directory in the package source code.

Running the Examples

To run the provided examples:

Ensure you're in the root directory of the project where package.json is located.
Install the required dependencies if you haven't already:
```
npm install
```
Run the example:
```
npm run example
```

Please note that you need to replace <YOUR_API_KEY> in the example script with your actual Talkscriber API key before running it.

The provided code is agnostic towards the sample rate and should be able to handle any .wav file/buffer that is pcm_s16le encoded.