AI Playground
Test and validate AI Review and AI Description behavior against real pull requests before applying changes. Playground provides a safe environment to refine rules, compare outputs, and ensure consistency across Git providers.
Playground lets LinearB Admins test, validate, and refine AI Review and AI Description behavior against real pull requests before applying changes to live workflows.
Use Playground to preview how LinearB AI responds to a selected pull request, test configuration updates safely, and confirm that your rules produce the expected output before saving them.
Accessing Playground
- In LinearB, open Automation from the left navigation.
- Select Playground (Beta).
- Select or enter a reference pull request from one of your connected repositories.
- Select the relevant Git provider integration.
Playground runs inside LinearB, so you can test AI behavior and review generated output without leaving the product.
What Playground helps you validate
Use Playground when you want to test how LinearB AI behaves before applying updates to your live configuration.
- Preview how AI Review comments on a reference pull request.
- Preview how AI Description generates pull request descriptions.
- Test custom guidelines before saving them.
- Validate rule or configuration changes against real pull request data.
- Compare results across different reference pull requests.
- Confirm behavior across supported Git providers.
Understanding the Playground interface
Playground provides a safe test environment for reviewing AI-powered automation behavior using a selected pull request.
- Pull Request selector — Select or enter the pull request URL you want to use as the test case.
- Integration selector — Select the connected Git provider integration you want to test.
- Configuration selector — Choose an existing configuration, such as a default provider configuration or a custom configuration.
- Configuration editor — Review or update the rule, guideline, or automation configuration before running the test.
- Validation status — Confirms whether the configuration syntax is valid before you run the test.
- Preview tab — Shows how the automation is expected to appear on the selected pull request.
- Output tab — Shows the generated result after running Playground.
- Context Variables tab — Shows available context data that can be used when validating automation behavior.
This allows you to test changes safely and understand how they affect real pull requests before applying them to live workflows.
Running Playground and reviewing results
- Select a pull request to use as the reference pull request.
- Select the Git provider integration.
- Select or edit the relevant configuration.
- Confirm that the configuration is valid.
- Run Playground.
- Review the generated result in the Preview or Output tab.
- Adjust the configuration and run Playground again if needed.
Playground simulates how the selected configuration behaves on the reference pull request. Changes are not applied to your live workflows until you save and deploy the configuration.
Recommended workflow
- Start with a representative pull request from the repository and provider you want to validate.
- Select the relevant integration and configuration.
- Add or update your custom guidelines for AI Review or AI Description.
- Run Playground and review the generated output.
- Refine the configuration until the output matches your expectations.
- Test against additional pull requests if the rule should work across different code patterns or repositories.
- Save your changes only after the Playground result looks correct.
Supported Git providers
Playground supports reference pull requests from repositories connected through the following Git providers:
- GitHub
- GitLab
- Bitbucket
Only repositories that are already connected and synced in LinearB are available for selection in Playground.
Requirements
- LinearB role: Admin.
- Required setup: A connected and synced Git provider integration.
- Reference pull request access: The selected pull request must be available to LinearB through the connected integration.
- Supported provider: The pull request must belong to a repository connected through GitHub, GitLab, or Bitbucket.
Troubleshooting
If a repository, integration, or pull request does not appear in Playground:
- Verify that the Git provider integration is connected and healthy.
- Confirm that the repository is synced in LinearB.
- Confirm that the pull request exists in a supported provider.
- Confirm that LinearB has access to the repository through the connected integration.
- Check whether the selected pull request URL is valid and accessible.
If Playground does not return the expected output:
- Confirm that the selected configuration is valid.
- Check the validation status before running the test.
- Review the selected pull request and confirm that it contains the type of changes your rule is intended to evaluate.
- Test with a different pull request to determine whether the behavior is specific to one pull request.
- Refine the configuration and run Playground again.
Known limitations
- Playground is available only for supported Git providers enabled for your organization.
- Only connected and synced repositories can be used.
- Playground output depends on the selected pull request, integration, and configuration.
- Playground is a validation environment and does not replace testing changes in your live workflow after configuration is saved.
Need help validating Playground behavior or confirming provider support? Contact support@linearb.io or visit the LinearB Help Center.
How did we do?
AI Iteration Summary for Teams
AI Rule Files