		@@ -23,84 +23,88 @@ ## Installing the module:

		```javascript
		const assemblyai = require('assemblyai')
		assemblyai.setAPIKey() // Note that the value is loaded from an environment variable.
		```
		### Initialization

		### Transcript
		The initialization of the module of course has to be at the beginning of your project.

		Transcribing audio from a URL

		```javascript
		const transcript = new assemblyai.Transcript()
		const assemblyai = require('assemblyai')
		assemblyai.setAPIKey("ENTER YOUR KEY HERE")
		```

		#### Creating a job
		### Upload an audio file for transcription

		When creating a job, you are required to pass in the audio source.

		An example with all of the options available is below, but the only required param would be `audio_src_url`.

		```javascript
		const response = await transcript.create({
		audio_src_url: "https://example.com/example.wav",
		model_id: 123,
		options: {
		format_text: true \|\| false
		async function upload () {
		try {
		const transcript = new assemblyai.Upload('/path/to/audiofile.wav')
		const response = await transcript.create()
		const data = response.get()

		// do something with the JSON response
		console.log(data);

		} catch (e) {
		// Do some error handling here
		}
		})
		}
		```

		#### Polling until job is complete

		Now that we have created a job, we can start polling for it.
		### Transcribe audio from a URL

		```javascript
		const { id } = response.get()
		const complete = await transcript.poll(id)
		```
		The only required parameter is the `audio_src_url` parameter. For more information about transcribing audio, please see the full API documentation [here](https://docs.assemblyai.com/api/#posttranscript).

		This ^ will wait until the job is complete or errors out until returning.
		If the job were to error out, an exception would be thrown.

		### Models

		Create a custom model to boost accuracy for phrases/keywords, or to add custom vocabulary.

		```javascript
		const model = new assemblyai.Model()
		async function transcribe () {
		try {
		const transcript = new assemblyai.Transcript()
		const response = await transcript.create({
		audio_src_url: "https://example.com/example.wav",
		model_id: 123,
		options: {
		format_text: true \|\| false
		}
		})
		const { id } = response.get()
		const data = await transcript.poll(id)

		// do something with the response data.
		// `data` is a wrapper of the API's JSON
		// response. `data.get()` returns the JSON
		// response of the API
		var responseJson = data.get();
		console.log(responseJson);

		} catch (e) {
		// Do some error handling here
		}
		}
		```

		#### Creating A Model

		Creating a model is very simple and is almost identical to creating a job.

		```javascript
		const response = await model.create({
		name: "foobar",
		phrases: ["foo", "bar"]
		})
		```

		#### Polling until the model has finished training
		### Create a custom model

		This API is identical to the Transcript method which is used like so:
		Boost accuracy for keywords/phrases, and add custom terms to the vocabulary with a custom model. For more information, please see the full API documentation [here](https://docs.assemblyai.com/guides/custom_models_101/).

		```javascript
		const { id } = response.get()
		await model.poll(id)
		async function model() {
		try {
		const instance = new assemblyai.Model()
		const response = await instance.create({
		phrases: ['foo', 'bar']
		})
		const { id } = response.get()
		const data = await instance.poll(id)

		// do something with the response data.
		// `data` is a wrapper of the API's JSON
		// response. `data.get()` returns the JSON
		// response of the API
		var responseJson = data.get();
		console.log(responseJson);

		} catch (e) {
		// Do some error handling
		}
		}
		```

		### Uploading audio files for transcription

		```javascript
		const upload = new assemblyai.Upload('/path/to/some/file.wav')
		```

		```javascript
		const response = await upload.create()
		const { text } = response.get()
		console.log(text)
		```

		## The Response Object
		### The Response Object

		@@ -113,2 +117,2 @@ When using the `Response` object, you will find a couple of methods:

		The method that you will most likely want to use will be `get()` which returns the full JSON object from the API, one level down.
		The method that you will most likely want to use will be `get()` which returns the full JSON object from the API, one level down.

src/api/util.js

		@@ -21,2 +21,6 @@ const Request = require('./Http/Request')
		}
		if (json.status === 'trained') {
		clearInterval(interval)
		resolve(response)
		}
		if (json.status === 'error') {
		@@ -23,0 +27,0 @@ clearInterval(interval)

assemblyai - npm Package Compare versions

New alerts

Fixed alerts

Improved metrics