Using the app is fairly intuitive. Just go on with your pull-request based workflow as usual.
When opening a pull request
Upon opening a pull request, you are greeted with the following comment from pull-dog where you can see that Pull Dog is provisioning a server for you.
After a brief while, the comment is updated with the connection details of your server. In the example on the screenshot, a docker-compose.yml was used which only exposes one port - 8000.
Closing a pull request
After closing the pull request, the server is shut down again, and the comment is updated to reflect that.