Skip to main content
Proton v5

Your first automation

This walkthrough takes an automation from scratch to its first run. Each step points to the page with the details.

Before you start​

  • The code. A git repository with the automation code, in the Proton format. The fastest way is to start from a template project, in Python or Java.
  • A runner. A machine with Proton Runner connected to the organization. See Environment Setup.
  • Permissions. A profile that can register systems, components, and automations, and run them in the Automation Center.
With AI

Proton MCP does this entire walkthrough from Claude or ChatGPT: it creates the project from the template, registers everything in Proton, and starts the first run.

1. Connect the git repository​

  1. In Settings > Integrations, use Create Integration and choose the repository platform (GitHub, GitLab, or Bitbucket), with the credential that gives access to it. Test Connection checks it.
  2. In Settings > Git Repository, use Create Git Repository: the platform, the Owner, the repository name, and the default branch.

2. Register the system​

In Settings > System, use Create System and link the system to the repository. Create one system per test platform: for example, the portal and the API, both linked to the same repository.

3. Register the components​

In Automations > Component, use New component for each step of the automation, in its system, with:

  • the name matching the component name in the code (the file in Python, the class in Java);
  • the input (in_) and output (out_) parameters that the code reads and returns;
  • the documentation: what it does, the goal, and the expected result.

See Components and Parameters.

4. Create the automation and the dataset​

  1. In Automations > Automations, use Add and give the automation a name. See Automations.
  2. Open the automation's Datasets and use Add.
  3. In the dataset, use Add Component to include the steps, in order.
  4. In Edit parameters, fill in the values for each step. Passwords and tokens are stored encrypted.

See Datasets and Building the dataset.

5. Run​

  1. In Automations > Automation Center, use Add to create an automation group and select the automation.
  2. In the group, select the dataset and use Run.
  3. Choose where it runs (the runner or a cluster) and Now, then confirm.

See Groups and cycles and Running.

6. See the result​

The run appears in the group with its live status. When you open it, the Timeline shows each step; the Results tab has the full log and the attachments; if the run fails, the Error tab has the AI diagnosis. See Run details.

The first run of a repository on a runner downloads the code and prepares the dependencies before the first step, and it can take a few minutes.

Next steps​

  • Create more datasets for other sets of data.
  • Schedule a run or create a recurring run in the Schedule.
  • Use an environment to run the same automation in staging and in production.