Add ScrapeUnblocker integration

[Kontuzijus]

Adds an integration page for ScrapeUnblocker.

ScrapeUnblocker renders web pages in a real browser behind anti-bot protections
(Cloudflare, DataDome, PerimeterX, Akamai) and returns raw HTML or AI-parsed
structured JSON, so a pipeline can ingest pages that a plain HTTP fetch cannot
reach.

The scrapeunblocker-haystack package is published and provides two components:

• ScrapeUnblockerFetcher - fetch URLs, one Document per page

• ScrapeUnblockerWebSearch - Google organic results as Documents

• PyPI: https://pypi.org/project/scrapeunblocker-haystack

• Repo: https://github.com/ScrapeUnblocker/scrapeunblocker-haystack

Both code examples in the page were run against the published package before
opening this PR. The logo is added as logos/scrapeunblocker.png.

[vercel]

@Kontuzijus is attempting to deploy a commit to the deepset Team on Vercel.

A member of the Team first needs to authorize it.

[Copilot]

Pull request overview

Adds a new Haystack integrations page documenting the scrapeunblocker-haystack package, so users can discover and use ScrapeUnblocker components in Haystack 2.x pipelines.

Changes:

• Adds integration front-matter metadata (logo, links, categorization) for ScrapeUnblocker.
• Documents installation and basic usage for ScrapeUnblockerFetcher and ScrapeUnblockerWebSearch.
• Adds an end-to-end pipeline example showing how to fetch a protected page and prompt an LLM with its contents.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[kacperlukawski]

Thank you, @Kontuzijus Are you sure the provided examples work, like pointed out by Copilot?

[Kontuzijus]

Good catch, and thanks for checking - Copilot was right, the pipeline example was broken.

ScrapeUnblockerFetcher outputs List[Document], while HTMLToDocument.sources expects list[str | Path | ByteStream], so that pipeline never even got to run - it failed at connect time:

PipelineConnectError: Cannot connect 'fetcher.documents' with 'converter.sources':
their declared input and output types do not match.
'fetcher':
 - documents: List[Document]
'converter':
 - sources: list[str | Path | ByteStream] (available)

Fixed exactly as Copilot suggested: the converter is gone and fetcher.documents now connects straight to prompt_builder.documents, since the fetcher already emits Document objects.

To make sure I am not guessing a second time, I executed every Python snippet on the page, parsed straight out of the updated scrapeunblocker.md, in a clean venv (haystack-ai 3.0.0, scrapeunblocker-haystack installed from PyPI) against the live API:

1. ScrapeUnblockerFetcher basic example - 1 Document returned, HTML content, meta url / content_type / parsed_data.
2. ScrapeUnblockerWebSearch basic example - 5 Documents with title / link / position in meta.
3. The pipeline example - runs end to end, including OpenAIChatGenerator, and answers the question about the fetched page.

One more docs fix in the same commit: base_url and timeout were missing from the ScrapeUnblockerWebSearch parameter table, although the component accepts both.