Google Cloud Speech API Node.js

The Cloud Speech API enables easy integration of Google speech recognition technologies into developer applications. Send audio and receive a text transcription from the Cloud Speech API service.

Table of Contents

• Setup
• Samples
  • Speech Recognition
• Running the tests

Setup

1. Read Prerequisites and How to run a sample first.

2. Install dependencies:

   With npm:

   npm install

   With yarn:

   yarn install

Samples

Speech Recognition

View the documentation or the source code.

Usage: node recognize.js --help

Commands:
  sync <filename>           Detects speech in a local audio file.
  sync-gcs <gcsUri>         Detects speech in an audio file located in a Google Cloud Storage bucket.
  sync-words <filename>     Detects speech in a local audio file with word time offset.
  async <filename>          Creates a job to detect speech in a local audio file, and waits for the job to complete.
  async-gcs <gcsUri>        Creates a job to detect speech in an audio file located in a Google Cloud Storage bucket,
                            and waits for the job to complete.
  async-gcs-words <gcsUri>  Creates a job to detect speech  with word time offset in an audio file located in a Google
                            Cloud Storage bucket, and waits for the job to complete.
  stream <filename>         Detects speech in a local audio file by streaming it to the Speech API.
  listen                    Detects speech in a microphone input stream. This command requires that you have SoX
                            installed and available in your $PATH. See
                            https://www.npmjs.com/package/node-record-lpcm16#dependencies

Options:
  --help                 Show help                                                                             [boolean]
  --encoding, -e                                                                          [string] [default: "LINEAR16"]
  --sampleRateHertz, -r                                                                        [number] [default: 16000]
  --languageCode, -l                                                                         [string] [default: "en-US"]

Examples:
  node recognize.js sync ./resources/audio.raw -e LINEAR16 -r 16000
  node recognize.js async-gcs gs://gcs-test-data/vr.flac -e FLAC -r 16000
  node recognize.js stream ./resources/audio.raw  -e LINEAR16 -r 16000
  node recognize.js listen

For more information, see https://cloud.google.com/speech/docs

Running the tests

1. Set the GCLOUD_PROJECT and GOOGLE_APPLICATION_CREDENTIALS environment variables.

2. Run the tests:

   With npm:

   npm test

   With yarn:

   yarn test