Demo App for Mozillas DeepSpeech project
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.
package.json feat(core): initial commit May 29, 2018
server.js feat(core): initial commit May 29, 2018

DeepSpeech Demo



This is a simple web-interface and Node server that uses DeepSpeech to create a local Speech-to-Text service.

I mostly built it for myself, in order to play around with DeepSpeech, but it can be used as a starting point for your own experiments.


The demo relies on SoX being available on your system and in your PATH.

A quick seach for how to install SoX on (windows|ubuntu|mac) should do the trick for you if you don't know how to install it, at least it did for me.


First, clone the repository:

git clone

Then change into the directory & perform an npm install:

npm install

As this downloads the pre-trained models from the DeepSpeech releases as a postinstall step, which are roughly 1.4 GB, it might take a while.

After that, start the server using:

npm start

And navigate to http://localhost:3000


The user interface is quite reduced, after you´ve agreed that the page is allowed to use your microphone, you just need to hit the "Listen" button and speak a few words into your microphone. After you finished your sentence, you need to hit the "Listening..." button once again, so that the recorded sound can be processed. After a few seconds, you should see the text in the result box below the button.