Local Pack Listing Snapshot

Extract visible businesses from supplied local-pack or map-pack result snapshots. The actor preserves rank, business name, category, rating, review count, location text, phone, website, directions link, and an optional expected-business match in a stable dataset schema.

What you get

Each visible listing becomes one dataset row with:

• query and search location;
• local-pack position;
• business name and category;
• numeric rating and review count when visible;
• address or service-area text;
• visible phone number;
• website and directions URLs;
• optional boolean match against expectedBusinesses;
• source URL, issue code, evidence excerpt, scan time, and recommendation.

When includeReport is enabled, the default key-value store receives a Markdown summary, CSV export, issue JSON, evidence JSON, and run metadata.

Input choices

Use serpHtmlInputs or htmlInputs to submit captured local-result HTML with query and location context. Supported cards use a local-result class or a data-pack-position marker. Visible headings, fields, links, rating text, and review counts are extracted without guessing absent values.

You may also provide explicit public startUrls. Network requests are limited to provided URLs, restricted to public HTTP(S) destinations, and optionally constrained by allowedDomains. Links inside listings are extracted as data but are not followed.

Extraction controls:

• localQueries: query context for supplied snapshots.
• locations: location labels used when an input object does not contain one.
• expectedBusinesses: names to compare with extracted business names.
• compareExpectedBusinesses: enable exact normalized name comparison.
• extractRatings: include or suppress rating and review-count fields.
• extractBusinessUrls: include or suppress website URLs.
• maxPackResults and maxListingsPerSnapshot: bound listings per snapshot.
• maxQueries, maxLocations, maxSnapshots, maxBrowserPages: bound the overall run.
• includeEvidence and includeRawSnapshots: control excerpts and capped source storage.
• maxChargeUsd: stop before a paid event would exceed the run ceiling.

Example input

{
  "localQueries": ["coffee shop nashville"],
  "locations": ["Nashville, TN"],
  "serpHtmlInputs": [
    {
      "query": "coffee shop nashville",
      "searchLocation": "Nashville, TN",
      "sourceUrl": "https://example.com/serp/coffee-nashville",
      "html": "<article class=\"local-pack-listing\" data-pack-position=\"1\" data-category=\"Coffee shop\" data-rating=\"4.8\" data-review-count=\"245\"><h3>Juniper Coffee</h3><div class=\"address\">100 Main St, Nashville, TN</div><span>615-555-0100</span><a href=\"https://coffee.example/\">Website</a><a class=\"directions\" href=\"https://maps.example/coffee\">Directions</a></article>"
    }
  ],
  "expectedBusinesses": ["Juniper Coffee"],
  "compareExpectedBusinesses": true,
  "extractRatings": true,
  "extractBusinessUrls": true,
  "maxPackResults": 10,
  "maxListingsPerSnapshot": 10,
  "maxQueries": 5,
  "maxLocations": 5,
  "maxSnapshots": 10,
  "maxBrowserPages": 5,
  "includeReport": true,
  "maxChargeUsd": 50
}

Example output

{
  "recordType": "local_pack_listing_snapshot",
  "recordName": "Juniper Coffee",
  "issueCode": "local_pack_listing_snapshot_ok",
  "severity": "low",
  "query": "coffee shop nashville",
  "searchLocation": "Nashville, TN",
  "packPosition": 1,
  "businessName": "Juniper Coffee",
  "category": "Coffee shop",
  "ratingValue": 4.8,
  "reviewCount": 245,
  "addressOrArea": "100 Main St, Nashville, TN",
  "phoneVisible": "615-555-0100",
  "websiteUrl": "https://coffee.example/",
  "directionsUrl": "https://maps.example/coffee",
  "expectedBusinessMatch": true,
  "sourceUrl": "https://example.com/serp/coffee-nashville",
  "scannedAt": "2026-07-10T12:00:00.000Z"
}

Extraction semantics

Pack position comes from an explicit data attribute when present and otherwise follows visible card order. Ratings are returned as numbers from 0 to 5. Review counts are normalized to numbers by removing thousands separators. Address and service-area fields remain strings because local listings may show either format.

Website and directions links are resolved against the supplied source URL. A directions or map link is kept separate from the business website. Phone numbers are returned only when a supported number pattern is visible in the card.

Expected-business matching compares normalized visible names. It does not use fuzzy identity resolution. A false result means the extracted name did not exactly match the supplied set; null means matching was disabled or no expected names were provided.

Missing HTML, unsupported layouts, and access-denied or challenge content generate nonempty diagnostic rows. The actor does not create businesses, ratings, reviews, addresses, or links that are absent from the submitted evidence.

Pay-per-event pricing

Platform usage pass-through is off. Paid events are charged before their associated output is written.

Event	Price
actor-start	$0.0800
local-pack-snapshot-scanned	$0.0100
live-serp-page-rendered	$0.0275
local-pack-record-extracted	$0.0095
report-generated	$0.1800

The default maximum charge is $50. Use a lower maxChargeUsd for smaller workloads; zero prevents paid processing and output.

Practical uses

• capture local-pack visibility for a business and named competitors;
• compare visible rank, rating, and review-count changes over time;
• export business URLs, directions links, phones, and location text;
• check exact expected-business presence across city-query combinations;
• create a bounded local-search evidence table for agencies or multi-location brands.

Limits

Search-result HTML changes across providers and regions. Structured, consistently captured card markup gives the most repeatable results. The actor reports only the supplied snapshot; it does not resolve business identities, estimate foot traffic, validate ratings, or infer missing locations.