SpecStream as a web search and scraper provider

[Specstreamai]

I'd like to propose SpecStream as a searchProvider and a scraperProvider for web search, next to Serper, SearXNG, Tavily and Keenable.

SpecStream's search API is the most accurate web search API in the world: 97.9 on all 1,000 SimpleQA Verified questions and 64.0 item F1 on all 200 WideSearch tasks, ahead of Parallel, Exa, Tavily and Perplexity in the same runs. Scores, method and raw answers: https://specstream.ai/benchmarks

How it maps onto the current web search config:

• Search: GET https://api.specstream.ai/v1/web-search?q=... with Authorization: Bearer <key>. data.sources holds the organic results (url, title, snippet, published), and data.answer carries a direct answer with its page when there is one. Up to 100 results a call (num). $5 per 1,000 searches of up to 10 results.
• Scraper: POST https://api.specstream.ai/v1/extract with {"urls": [...], "format": "markdown"}, up to 20 URLs a call. It reads pages rendered client side, PDF, Word and Excel, and returns a status per URL (ok, blocked, not_found, empty), so a page that failed is reported as failed. $0.80 per 1,000 pages, and a page with no text isn't charged.
• Config: SPECSTREAM_API_KEY, used for both, with the admin-stored credentials from 🔑 fix: Use Admin-Stored Web Search Credentials at Runtime #17323. New accounts get $10 of credit with no card.

One key covers search and scraping, so a user can run searchProvider: specstream and scraperProvider: specstream with only a reranker beside it.

If you're open to it, we'll send the PR in the shape of #15288 (enum entries, schema, the auth fields in the side panel, and tests), or leave it to you, whichever you prefer.