Local Review Snippet Extractor

Extract visible review snippets, ratings, dates, reviewer display names, response signals, and transparent text themes from public business pages.

What you get

• Fetch only supplied public review or business page URLs; arbitrary review-site crawling is disabled.
• Extract repeated visible review blocks and JSON-LD Review entities.
• Normalize ratings, publication dates, reviewer display names, owner-response signals, and source selectors.
• Filter by visible minimum rating and cap snippets per page and source count.
• Assign transparent keyword themes such as service, speed, quality, price, and communication.

Input

{
  "startUrls": [
    "https://example.com/business/reviews"
  ],
  "reviewPageUrls": [],
  "allowedDomains": [
    "example.com"
  ],
  "maxPages": 3,
  "maxDepth": 0,
  "maxSources": 3,
  "businessName": "Example Plumbing",
  "minRating": 1,
  "maxReviewsPerPage": 10,
  "extractThemes": true,
  "includeReviewerDisplayNames": true,
  "includeEvidence": true,
  "includeRawSnapshots": false,
  "requestDelayMs": 250,
  "timeoutMs": 15000,
  "userAgentMode": "standard",
  "maxChargeUsd": 15,
  "includeReport": true
}

Use htmlInputs when you already have source HTML or need deterministic recurring analysis. URL runs accept normal public HTTP(S) pages. Private-network and unsupported URL targets are rejected.

Dataset output

Each dataset item is a complete, schema-defined record. Fields that are not present in the source remain null rather than being guessed.

{
  "sourceUrl": "https://example.com/business/reviews",
  "recordType": "review_snippet",
  "issueCode": "local_review_snippet_extractor_review_extracted",
  "reviewSource": "example.com",
  "businessName": "Example Plumbing",
  "ratingValue": 5,
  "ratingScale": 5,
  "reviewDate": "2026-06-01",
  "reviewerDisplayName": "Taylor",
  "reviewText": "Fast service and clear communication.",
  "reviewTheme": "service",
  "sentimentLabel": "positive",
  "responseFound": true,
  "evidenceSelector": "review-block:1"
}

Reports

A successful run can write:

• local-review-snippet-extractor-summary-report.md
• local-review-snippet-extractor-records.csv
• local-review-snippet-extractor-issues.json
• local-review-snippet-extractor-evidence-samples.json
• local-review-snippet-extractor-run-metadata.json

Reports include row counts, issue severities, evidence references, recommendations, and the input fields consumed by the run.

Pricing and cost controls

This actor uses pay-per-event pricing with Apify platform usage pass-through off:

• actor-start — $0.01000 per run
• review-source-page-scanned — $0.00650 per page
• review-snippet-extracted — $0.00295 per snippet
• review-theme-emitted — $0.00295 per theme
• audit-report-generated — $0.08000 per report

The default maxChargeUsd is $15. The runtime checks the local budget and the Apify charge result before every paid dataset item. If either limit is reached, it stops before emitting unpaid output and writes billing diagnostics to key-value storage.

Safe crawling behavior

• Public HTTP(S) sources only.
• Private-network, localhost, and unsupported protocol targets are rejected.
• allowedDomains, maxPages, actor-specific caps, maxDepth, timeoutMs, and requestDelayMs bound discovery.
• Supplied HTML is processed without an external request.
• Protected or authenticated sources are outside the supported public-data scope.

Limitations

• Only publicly visible review text is extracted; no accounts, private profiles, or hidden reviewer data are accessed.
• Missing ratings, dates, and display names remain null rather than being inferred.
• Keyword themes are transparent evidence labels, not model-generated sentiment claims.

Common use cases

• Export source-linked review evidence.
• Find negative reviews without a visible owner response.
• Compare visible ratings and themes across supplied sources.