[FEATURE] A web-scraping tool whose browser session survives between runs

Feature Area

Integration with external tools

Is your feature request related to a an existing bug? Please link it here.

NA

Describe the solution you'd like

Every browser tool under web-scraping today starts from nothing on each call: a fresh browser, a stock automation fingerprint, and whatever IP the process has. That is fine for a public page and useless behind a login, because the session dies with the browser, so a crew that has to read a signed-in page cannot keep its state between steps.

I would like a tool whose unit of state is a browser profile rather than a session: cookies, storage, a fingerprint and a proxy that all persist, so the second call opens the page already signed in, and one browser is reused across the calls of a single crew run.

Concretely, a AntibrowLoadTool(profile="research-01") that takes a url (and an optional CSS selector), returns the page's visible text truncated to a limit, and exposes close() for cleanup.

Describe alternatives you've considered

• SeleniumScrapingTool with a user-data-dir: keeps cookies, but the fingerprint is a stock automation build and the setup is left to the user.
• The hosted-browser tools (BrowserbaseLoadTool, HyperbrowserLoadTool): sessions are remote and per-call, and they bill by browser time.
• Doing it outside the crew and passing text in: works, but then the login step is not something an agent can drive.

Additional Context

I am affiliated with AntiBrow, which is why I can implement it: a PR is ready with the tool, 9 unit tests, the optional extra, and the docs page in en/ar/ko/pt-BR.

Start by reviewing the existing browser tools under web-scraping and the proposed AntibrowLoadTool(profile="research-01") interface. Then inspect the nine unit tests, optional extra, and documentation pages in en/ar/ko/pt-BR mentioned in the issue. Done means the profile state and browser reuse behavior work across calls, cleanup is exposed, and the documented interface is covered.