SERP Feature Snapshot Extractor

Detect and extract visible SERP features such as featured snippets, image packs, video blocks, local packs, shopping blocks, knowledge panels, top...

Who it is for

SERP Feature Snapshot Extractor helps seo teams, content strategists, serp feature analysts, agencies detect and extract visible SERP features such as featured snippets, image packs, video blocks, local packs, shopping blocks, knowledge panels, top stories, and result modules. It produces schema-valid dataset rows, evidence-backed KVS reports, safe diagnostics, and capped pay-per-event pricing for predictable Apify Store runs.

Inventory visible featured snippets, image packs, video modules, local packs, shopping blocks, knowledge panels, and top-stories modules. Each row preserves the visible module text, owner domain when linked, position, item count, and source snapshot key.

Evidence contract

The actor emits only facts visible in supplied HTML, explicit input records, or a permitted public URL response. It never creates rankings, questions, modules, suggestions, titles, snippets, or page metadata to fill an evidence gap. A missing source produces a schema-valid status row explaining what input is needed.

Network access is conservative. Only HTTP and HTTPS URLs are accepted; localhost, loopback addresses, private IPv4 ranges, and disallowed domains are rejected. The runtime does not follow links discovered in a page. Challenge and access-denied pages are recorded as blocked evidence and are not solved or circumvented.

Input modes

• Pass serpHtmlInputs or htmlInputs with captured search-page HTML.
• Use featureTypes to limit output to normalized module types such as featured_snippet or local_pack.
• Use startUrls only for public URLs that return the search-page HTML you are authorized to process.
• Set includeModuleScreenshots to receive a clear snapshot-reference note; this runtime records HTML evidence and does not create image files.

Example input

{
  "queries": [
    "best crm"
  ],
  "serpHtmlInputs": [
    {
      "sourceUrl": "https://example.com/serp",
      "query": "best crm",
      "html": "<section class=\"featured-snippet\" data-position=\"1\"><h2>What is a CRM?</h2><p>A CRM organizes customer relationships.</p><a href=\"https://vendor.example/guide\">Source</a></section>"
    }
  ],
  "featureTypes": [
    "featured_snippet"
  ],
  "maxFeaturesPerSnapshot": 25,
  "maxChargeUsd": 50
}

Dataset output

Each dataset item is one JSON object, which makes the result directly usable in Apify integrations, exports, webhooks, and downstream data pipelines.

• featureType and featurePosition
• visible feature title, text, evidence excerpt, and item count
• first visible feature URL and owner domain when present
• query, source status, and stable snapshot key

Common evidence fields include sourceUrl, sourceType, pageTitle, issueCode, severity, evidenceSnippet, recommendation, httpStatus, and scannedAt. Fields without source evidence are null rather than guessed.

Reports and retained evidence

When includeReport is true, the actor writes a Markdown summary, CSV records, JSON issue rows, evidence excerpts, and run metadata to key-value storage. Set includeRawSnapshots only when retaining capped HTML is appropriate for your workflow; it is off by default.

The report metadata records the actor, timestamp, row count, issue codes, severity totals, consumed input fields, pricing mode, platform-usage setting, and maximum-charge default. This provides a compact audit trail for scheduled comparisons.

Pricing

This actor uses pay per event. Apify platform-usage pass-through is off. Charges are accepted before the corresponding paid dataset row or report is written.

Event	Price
actor-start	$0.08000
query-snapshot-scanned	$0.01000
live-serp-page-rendered	$0.02750
serp-feature-record-extracted	$0.00950
report-generated	$0.18000

maxChargeUsd is a strict pre-output budget. A value of zero prevents the first charge and therefore prevents paid output. If Apify rejects a charge, partially accepts it, or reports an event limit, the actor stops before writing the affected output and records billing status.

Higher-tier Store discounts are expected where supported. The event prices above remain the locked base prices.

Limits and predictable runs

• maxPages, maxSnapshots, and actor-specific record caps limit work before analysis.
• maxQueries limits distinct search contexts.
• Public fetches are additionally constrained by allowedDomains, timeoutMs, requestDelayMs, and a fetch cap.
• Reports are optional and raw HTML retention is opt-in.
• Query-only input produces an explicit source-required row instead of an empty or fabricated result.
• Dataset rows and report files follow the declared Actor schemas.

Responsible use

Process only public content or material you are authorized to analyze. Respect website terms, rate limits, privacy obligations, intellectual-property rights, and applicable law. Avoid sending personal or confidential information in HTML or input records unless your workflow is authorized to process it.