This code sends the speech to text output of the Amazon Echo across a ROS network on a topic.
To get this running, first you need to go into src/index.js and you need to change the APP_ID variable to your own app_id, and change the url variable to the address of your websocket server.
Then you'll need to install the roslibjs library with npm npm install roslib.
Then you'll need to zip the files and folders in src into a single zip, upload that to a lambda function with Amazon, and link that function to an Alexa skill. Check here for more detailed instructions.
After that, start your websocket server on your local machine with roslaunch rosbridge_server rosbridge_websocket.launch.
Now enable the skill though Amazon's web interface, and you're good to go! The topic name is /speech_recognition, but feel free to change that.