Proprietary runner versus standard code in AI testing tools

As of 2026-09-24, the catalogue documents test output for 156 of 330 tools. Of those 156, 105 run tests on a proprietary runner and 19 produce standard framework code. Output is not documented for 158 tools, which is more than any single documented value.

What the output fact records

The output fact records whether the tests a tool creates can run without the vendor's product. It takes one of five values from the taxonomy:

• proprietary-runner: as far as the docs show, tests run only inside the vendor's product.
• standard-playwright, standard-cypress, standard-selenium: the docs state that the tool generates or exports plain framework code that runs without the vendor.
• mixed: the docs describe both a vendor runner and standard framework code.

Mixed covers tools where tests run on the vendor's runner and an export to standard code also exists. For example, mabl tests run on mabl's runner. The mabl help article on exporting tests from the CLI lists export formats that include Playwright.

An export can have limits, such as test types it does not cover. Each tool page links the source page, where those limits are stated.

Distribution of documented output

Counts as of 2026-09-24, sorted by count. Shares are of the 156 tools with documented output, rounded to whole numbers.

Outside the documented group, output is vendor-stated for 16 tools. It is not documented for 158 tools.

Not documented means the public pages we opened do not say where tests run. It does not mean the output is proprietary. The post How to check test output ownership shows where to look in vendor docs.

Tools with documented standard Playwright output

Sorted by name A-Z. All 14 are documented and verified on 2026-09-24.

Five of the 14 are in the agentic end-to-end testing category. Three are in the MCP servers category.

The same list, with filters, is on the standard Playwright output feature page. The Cypress and Selenium tools are on the standard Cypress output and standard Selenium output pages.

How the lock-in filter uses this fact

The avoid lock-in page lists tools whose output fact is standard Playwright, Cypress or Selenium. Documented matches are in the main table. Vendor-stated matches are listed separately, with their source and date.

The page does not include mixed tools. A mixed tool documents a vendor runner, so its tests may not all run without the vendor. The tool page and its source show which tests the export covers.

Sources

• mabl help: Exporting tests from the CLI
• AISquare Studio QA repository
• Bugzy (Felan) docs: QA use case
• Checkly docs: Playwright check construct
• Checksum docs: Story and Test Format
• ExecuteAutomation Playwright MCP docs: Recording actions
• Octomind docs FAQ
• Ogoron docs: Test execution
• Playwright CLI docs: codegen
• Playwright MCP repository
• Playwright docs: Test agents
• QA DNA pricing
• Qodex docs: Scripts
• Quorvex AI repository
• RoostGPT docs: UI tests in CI/CD
• Tool pages: mabl, All tools