Add Microsoft Web IQ as a search and scraper provider

[mainhanu]

Summary

Adds microsoftWebIQ as a new option for both the search provider and the scraper provider in the search tool, alongside the existing Serper / SearXNG / Tavily / Firecrawl integrations.

core logic is src/tools/search/microsoft-scraper.ts & src/tools/search/microsoft-search.ts(~ 500 lines), most of the rest code is for ut & local test script

What's included

• Search API (microsoft-search.ts) — wraps the Web IQ search endpoints:
  • POST /v3/search/web (with contentFormat: 'passage' for query-relevant snippets) → organic.
  • POST /v3/search/news → topStories + news, mirroring how Serper folds news into top stories.
  • POST /v3/search/videos → videos (title, link, snippet, thumbnail, duration, channel, source, date).
  • POST /v3/search/images → images (title, image url, dimensions, thumbnail, host page link, source/domain).
  • safeSearch is mapped to the API's off / strict enum for the videos and images endpoints.
• Browse scraper (microsoft-scraper.ts) — implements BaseScraper over POST /v3/browse with liveCrawl: 'none' (indexed content only). Non-200 responses (202 crawl-in-progress, 404 not-indexed) are treated as scrape failures so the source is skipped rather than retried.
• Wiring — registers the provider in createSearchAPI and the scraper selector in tool.ts, including config plumbing (microsoftWebIQApiKey, microsoftWebIQBaseUrl, microsoftWebIQSearchOptions, microsoftScraperOptions). Videos and images are supported, so the parallel type: 'videos' / type: 'images' sub-searches fan out and merge into the result.
• Types — adds MicrosoftWebIQSearchOptions, MicrosoftScraperConfig, MicrosoftScrapeOptions, and MicrosoftBrowseResponse; extends the SearchProvider / ScraperProvider unions.
• Tests (microsoft.test.ts) — covers missing API key, empty query, web→organic, news→topStories, videos→videos, images→images mapping, request failure, and scraper 200/202/404 handling (11 tests).
• Smoke script (scripts/microsoft_search.ts, npm run search:microsoft) — exercises web, news, video, and image search, the Browse scraper, and the full search tool against the live API.

Configuration

Credentials are read from config or MICROSOFT_WEBIQ_API_KEY / MICROSOFT_WEBIQ_BASE_URL (defaults to https://api.microsoft.ai/v3).

Testing

• npx jest src/tools/search/microsoft.test.ts — 11/11 passing
• tsc --noEmit and ESLint clean
• Local linked to the LibreChat app, demo videos below (will raise PR when agent is merged and published).

search

librechat.search.mov

scraper

librechat.loader.-.1080p.mov