Web Search API with Page Markdown

Run web searches programmatically through a web search API and get structured results: rank, title, URL and snippet for every hit. Turn on one flag and the Actor also fetches each result page as clean Markdown — search plus content in a single run. Built for AI agents, RAG pipelines and research jobs that need real ranked hits, not a static URL list. When fetching pages, the Brainiall Web engine honors robots.txt, does not bypass bot detection, and refuses internal or private addresses. Unfetchable pages are omitted from the page event and not charged.

What it does

• Web search — submit one or more queries; receive up to 25 ranked hits each.
• Structured output — one dataset item per hit: query, rank, title, url, snippet.
• Optional page Markdown — with fetchPageContent enabled, each item also carries the page body as normalized Markdown, ready to chunk for retrieval.

Leave fetchPageContent false when you only need links. Turn it on when the pipeline needs the page body. Batch related queries in one run; they execute concurrently. Deduplicate by url when queries overlap.

Why this Actor

• Search plus Markdown in one run. Research agents get live ranked hits and optional page text without chaining a SERP tool and a scraper.
• Metered, exportable dataset. Ranked hits land as dataset items you can schedule, export as JSON/CSV/Excel, and charge per result — not a one-off HTML dump.
• Page fetch is opt-in. Keep it off for cheap link discovery at $0.001 per hit. Turn it on only when you need the body ($0.002 per converted page).

Input

Field	Type	Default
queries	array of string	— (required)
maxResultsPerQuery	integer	10 (maximum 25)
fetchPageContent	boolean	false

{
  "queries": [
    "best vector databases 2026",
    "retrieval augmented generation best practices"
  ],
  "maxResultsPerQuery": 10,
  "fetchPageContent": true
}

Output

{
  "query": "best vector databases 2026",
  "rank": 1,
  "title": "Best Vector Databases in 2026: Complete Comparison Guide",
  "url": "https://example.com/articles/best-vector-databases",
  "snippet": "Compare the best vector databases in 2026: features, performance, pricing...",
  "markdown": "# Best Vector Databases in 2026\n\n..."
}

Without fetchPageContent, markdown is absent and you are charged only for search results. Export JSON, CSV or Excel, or read the dataset through the Store API.

Pricing

Pay-per-event. Current prices (do not assume other SKUs):

Event	USD
search-result (primary)	$0.001 per ranked result delivered
page-converted	$0.002 per result page converted to Markdown (only when fetchPageContent is on)

Worked cost example:

• 10 results, links only → $0.01
• 10 results + 10 pages fetched → $0.01 + $0.02 = $0.03
• 3 queries × 10 results, no pages → 30 × $0.001 = $0.03

Failed queries and unfetchable pages are never charged. No subscription and no monthly minimum. Set BRAINIALL_API_KEY as an Actor secret (key from https://app.brainiall.com). Keys are never shipped in the listing.

Use cases

1. Research agents that need live search plus page text in one call.
2. RAG pipelines that turn a topic into fresh Markdown documents from current ranked pages.
3. SEO and content research that tracks what ranks for a query and what those pages actually say.

FAQ

Do I pay for pages I did not ask for? No. page-converted fires only when fetchPageContent is true and the page actually converts.

Can I get 100 results per query? No. The cap is 25 per query. Split into several focused queries.

Is this a full-site crawler? No. Use Markdown Website Content Scraper when you already have URLs or a site root. This Actor starts from search queries.

Limits

• maxResultsPerQuery: 1–25 (hard cap).
• Page fetch is opt-in. Keep it off for cheap link discovery.
• robots.txt and bot detection apply to page fetches. Blocked pages skip the page-converted event.
• Internal, loopback and link-local URLs are never fetched.
• Batch related queries in one run; they execute concurrently. Deduplicate by url when queries overlap.

---

Built and maintained by Brainiall — production speech, document and web engines.