Skip to content

Step-by-Step Guide

DrupalPod AI QA is an automation tool designed to simplify testing and validation of AI module patches and pull requests in Drupal environments.

This guide covers using the DrupalPod AI QA Chrome extension with DrupalForge.

Prerequisites

Make sure you have:

DrupalPod AI QA helper extension in the Chrome Web Store

Step 1: Open the Ticket

Navigate to the AI module issue you want to test, then open the Chrome extension while you are on that ticket page.

AI module issue list on Drupal.org

Open the specific issue you want to test:

Example issue page on Drupal.org

If the issue has an issue fork, click Get push access before launching:

Get push access on the issue page

Step 2: Configure the Environment

Inside the extension, review the environment variables before launch.

Most users will not need to change anything. In the standard flow, the extension will populate the required values from the ticket context.

If you are not sure, leave the defaults alone and use the values generated by the extension.

Extension opened on the issue page

Scroll through the version settings and review the launch target:

Extension version settings and launch options

Step 3: Choose Where to Build

Select the target environment:

  • DrupalForge if you want a hosted environment
  • Local if you want to run the project in DDEV on your machine

Local DDEV option

The local option requires a manual setup step. See the Local DDEV Guide for details.

For the DrupalForge path, make sure you are already signed in to DrupalForge in another tab. The extension will use that session to launch the environment.

When you are ready, click the launch button in the extension:

Launch DrupalForge from the extension

Step 4: Review and Launch in DrupalForge

After clicking launch, DrupalForge will open a generated application screen.

AI provider selection

In the standard workflow, DrupalPod AI QA uses ai_provider_amazeeio, which gives you an auto-provisioned trial setup. You can optionally set DP_AI_PROVIDER to openai or anthropic and enter your own key. For more detail, see Choosing an AI Provider.

At this stage:

  1. Review the generated environment information
  2. Confirm the required terms and policy checkbox
  3. Click Launch your application
  4. Wait for the environment to finish provisioning

Generated environment information in DrupalForge

Check the required agreement box:

Terms and conditions checkbox before launch

If prompted, log in to DrupalForge or DevPanel:

DevPanel sign-in screen

Step 5: Wait for Deployment

After launch, the environment will enter deployment.

During deployment, you can monitor the status in DevPanel:

DevPanel deployment in progress

Step 6: Access the Environment

Once deployment is complete, the DrupalForge dashboard will provide access to:

  • The running site
  • DevPanel controls
  • An IDE for inspecting or editing the project code

Projects are expected to remain active for six hours.

After deployment is complete, the application summary and browser-based tools will be available from the DevPanel overview:

DevPanel overview after deployment

You can access your application by clicking the URL in the Application Summary. To log into admin, go to https://[your-site].devpanel.app/user/login and use admin:admin credentials to log in.

And then your application is ready to use! You can start testing your issue branch, or explore the codebase by using VS Code in the browser.

Public environment and AI provider credentials

DrupalPod AI QA environments should be treated as temporary shared environments and are using insecure admin:admin credentials for the super admin user.

If you want to use a provider that requires an API key, see Choosing an AI Provider for the DrupalPod AI QA temporary key flow and DP_AI_PROVIDER options. This will allow you to use your own key, and will auto-expire the key after a limited time window.

VS Code in Browser

If you want browser-based VS Code access, go to the Start VSCode In Browser section, copy the generated password, open the application, and submit the password in the prompt.

For the access steps, see Working with Visual Studio Code in the Browser.

If the site does not build correctly, see Troubleshooting.