After you have installed and configured the Pipelines client, which is required only if you want to manage Pipelines from the command prompt, complete the following steps to create an application with the command-line client:

1. Create a build definition file that describes how to build your application and add it to the root of your workspace.

2. Build your application using the start command.

   Note for GitHub users

   As an alternative, you can use the Cloud Platform Pipelines integration in the user interface, in which case creating a pull request, pushing a tag, or pushing a branch in GitHub triggers a Pipelines build.

   When you run the pipelines start command in a Git repository that has a Cloud Platform repository as a remote, the start command determines the correct application ID. However, if your local Git repository has only GitHub as a remote, the start command cannot determine the correct application ID on its own. You can set the correct application ID for a specific repository using the pipelines set-application-id command.

   To determine the correct application ID, run the pipelines list-applications command, and then find the application ID associated with the application that you want. Then, from the Git repository that you want to configure, run the following command:

   copy

   pipelines set-application-id --application-id=[application ID]

3. Use the status command to verify if the build has completed and whether it was successful. If you use the Cloud Platform Pipelines GitHub integration, you can see job completion and success information on the GitHub pull request. You still need to use the status command for information about commits that are not part of a pull request.

4. If desired, deploy your build artifact.