https://github.com/rxtoolkit/stt-gcp

👂 RxJS operator for realtime speech-to-text (STT/S2T) using GCP speech-to-text
https://github.com/rxtoolkit/stt-gcp

fp functional-programming observables package reactive-programming rxjs speech-to-text stt transcription

Last synced: 6 days ago
JSON representation

👂 RxJS operator for realtime speech-to-text (STT/S2T) using GCP speech-to-text

• Host: GitHub
• URL: https://github.com/rxtoolkit/stt-gcp
• Owner: rxtoolkit
• License: mit
• Created: 2021-02-09T21:25:55.000Z (almost 4 years ago)
• Default Branch: master
• Last Pushed: 2024-02-16T18:40:09.000Z (11 months ago)
• Last Synced: 2024-12-15T19:25:40.805Z (26 days ago)
• Topics: fp, functional-programming, observables, package, reactive-programming, rxjs, speech-to-text, stt, transcription
• Language: JavaScript
• Homepage:
• Size: 3.57 MB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 10
• Metadata Files:
  • Readme: README.md
  • Changelog: CHANGELOG.md
  • License: LICENSE.md

Awesome Lists containing this project

README

        
# @rxtk/stt-gcp

> 👂 An RxJS operator for real-time speech-to-text (STT/S2T) streaming using the GCP speech-to-text API.

> 🍎 If you are using an Apple Silicon (ARM) machine, this package only works in node >15.x

```bash

npm i @rxtk/stt-gcp

```

```bash

yarn add @rxtk/stt-gcp

```

⚠️  To run the GCP transcription pipeline, you will need to provide a path to JSON containing your `GOOGLE_APPLICATION_CREDENTIALS`. You can set this as an environment variable or pass it in.

⚠️  node.js only. This has not been tested on Browsers but it might be possible to make it work.  If you get it working, please make a PR!

## API

### `toGCPSpeech`

Stream audio speech data to AWS Transcribe via WebSocket and get transcripts back:

```js

import {map} from 'rxjs/operators';

import {toGCPSpeech} from '@rxtk/stt-gcp';

// The pipeline can take a stream of audio chunks encoded as 

// LINEAR16 (PCM encoded as 16-bit integers) in the form of a Buffer

const buffer$ = pcmChunkEncodedAs16BitIntegers$.pipe(

  map(chunk => Buffer.from(chunk, 'base64')),

  toGCPSpeech({

    sampleRate: 16000, // optional

    contentType: 'audio/l16', // optional

    googleCreds: process.env.GOOGLE_APPLICATION_CREDENTIALS, // required

    useSpeakerLabels: true, // optional

    useWordConfidence: true, // optional

    usePunctuation: true, // optional

  })

);

buffer$.subscribe(console.log); // log transcript output

```

## Guides

- [Introduction to audio data](https://developer.mozilla.org/en-US/docs/Web/Media/Formats/Audio_concepts)