[Feature]: Playwright Test Agents support for Python (planner / generator / healer)

[DepravityDemo]

🚀 Feature Request

Hi Playwright team,

First of all, thank you for releasing Playwright Test Agents (planner / generator / healer).
This is an amazing step toward AI-assisted test authoring and maintenance.

However, Playwright Test Agents currently support only the JS/TS Playwright Test runner.
Many teams, including ours, primarily use Playwright for Python, and we would really like to use the same agentic workflow:

planner → explore the app and produce a Markdown test plan
generator → convert the plan into Playwright Python test files
healer → execute pytest/unittest and automatically repair failing tests

Request
Would the team consider adding official Python support for Playwright Test Agents?

Specifically:

Agent definitions tailored for Python (planner / generator / healer)
Python test generation templates
A "healer" loop that integrates with pytest or python-based runners
playwright init-agents --loop=python (similar to existing vscode / claude / opencode loops)

Example

No response

Motivation

Playwright Python is widely adopted in QA automation (especially in enterprise environments).
Many Python teams want to leverage AI-driven test creation and self-healing.
Currently, the only way is to convert JS/TS output into Python manually or build custom agents.
Having first-class Python support would make the new agentic workflow accessible to a much larger community.

Thanks for considering this!

[m13v]

the planner→generator half ports to python fine: the plan is just markdown, generation is templating. the healer is where the taxonomy leaks. repairing a failing locator needs the runner's trace introspection, and that signal lives in the JS test runner, not in pytest's failure output. so --loop=python gets you two of the three agents almost for free, but the healer isn't a template swap, it's the real port. worth scoping that one separately from planner/generator.