Command Line Client

After building your model in the Speechly Dashboard, it’s time to start developing your application.

Once you’ve created your application in the Speechly Dashboard and tested it in the Playground to ensure it performs correctly, it’s time to put it into use.

The easiest way to see your model working in a real-life application is to use our simple command line client. You can find the repository in GitHub.


To begin with, you need a Mac OS or a Linux.

Developing on Windows?

If you are developing on Windows, you can install Linux on a virtual machine by following these instructions.

You will also need to have Node (version >= 12) and sox installed. For Mac OS you can do:

brew install sox node

And for most Linux distros:

sudo apt-get install nodejs sox libsox-fmt-all

Installing the client

Now you can install the client and move on to the correct directory by using:

git clone && cd start/client

Then, install the dependencies by using:

npm install

Using the client

Start the client by running:

node audio.js

The client will ask for your application ID. Provide the app ID from the Speechly Dashboard, and you’ll see the line:

--- Start new audio context ---

This means that the client is listening to the audio input. Then say something that your model understands, and the client will provide you with the transcript, intents, and entities in real-time, just like in the Playground.

Screenshot from the Speechly Command Line Client
The Speechly Command Line client.

You can start a new audio context by clicking the space bar. You can read more about the audio context here.

What next?

Now that you know your model works and the type of data Speechly returns, you can start the actual development. Although you can use the command line client as the scaffold for your application, easier yet is to use one of our client libraries.

Profile image for karoliina-louhema

Last updated by karoliina-louhema on May 26, 2020 at 09:16 +0300

Found an error on our documentation? Please file an issue or make a pull request