Skip to main content

Our Approach

Playmatic’s test execution engine combines self-healing capabilities with deterministic caching to minimize flakiness. During a test run, Playmatic will always execute any cache function first. If the cache function fails, then Playmatic will attempt to self-heal the step by using a computer-use agent. The agent will use your English instructions in the test step, test goal, and the cache function code during self-healing.

Playmatic's logic to determine test step failure

Executing Tests

Playmatic tests can be run locally in your CLI or remotely in the CI.

Locally

Run tests on your development machine using the Playmatic CLI:
# Run tests in the default test environment
playmatic run

# Run tests in a specific test environment
playmatic run --env production
You can run tests against a localhost or remote environment in the CLI.

In CI

Automate test execution in your continuous integration pipeline. Playmatic integrates with GitHub Actions to run tests automatically on merge, pull requests, or on a schedule.

Test Environments

Playmatic supports running tests in both local and remote application environments. Test environments can be setup in your playmatic.config.ts file, and selected at runtime. You can configure the following for each test environment:
  • baseUrl (string): The base URL for the application instance
  • vars (object, optional): Environment-specific variables accessible via env.vars in tests
playmatic.config.ts
export default {
  env: {
    staging: {
      baseUrl: "https://staging.playmatic.ai",
      vars: {
        TEST_USER_EMAIL: "james@playmatic.ai",
        TEST_USER_PASSWORD: "bugkiller123!"
      },
    },
  },
};

Parallelization

By default, all Playmatic tests are executed in parallel within a single run, with a maximum of 5 tests running simultaneously. If you have dependencies between the tests, we recommend combining the test steps into a single test or running them separately.

Debugging

Playmatic provides comprehensive debugging tools after each run. You get a complete report with visual playback and detailed traces to quickly identify and resolve issues.

Example of a test run with 3 tests

Next Steps