Yellow Pages Scraper — US local business data with phone, address and website

Scrape business listings from Yellow Pages by search term and city: name, phone, full address, categories, rating, review count, opening hours and website — 30 businesses per page, straight into a table.

Why this one holds up

Most Yellow Pages scrapers read the page's HTML classes. Yellow Pages changes those, and those scrapers quietly return empty rows until someone notices.

This one takes its core fields from the page's structured data — the LocalBusiness blocks the site publishes for search engines. Yellow Pages cannot change those without damaging its own SEO, which is exactly why they are stable.

The markup is still used, but only for the three things structured data leaves out — categories, website and position — and the two sources are joined by each business's own listing path rather than by row order. That matters more than it sounds: joining by position silently mismatches rows whenever the page escapes a character in a name, and joining by the site's internal ID only works for about half the listings. The listing path matched every single one.

---

Quick start

{ "searchTerms": ["plumber"], "location": "Austin, TX" }

Several trades at once, only established businesses:

{
  "searchTerms": ["plumber", "electrician", "hvac contractor"],
  "location": "Phoenix, AZ",
  "minRating": "4",
  "minReviews": 10,
  "maxPagesPerSearch": 10
}

Lead lists — businesses that have a phone but no website yet:

{
  "searchTerms": ["landscaping"],
  "location": "Denver, CO",
  "onlyWithPhone": true,
  "onlyWithWebsite": false
}

Or paste result pages you already filtered on the site:

{ "startUrls": ["https://www.yellowpages.com/austin-tx/plumbers?page=2"] }

---

Input

Field	Type	Default	Description
searchTerms	array	["plumber"]	Trade or business type to search
location	string	Austin, TX	City and state. Required with search terms — Yellow Pages never searches nationwide
startUrls	array	—	Result pages you already filtered on the site
maxPagesPerSearch	integer	5	30 businesses per page
minRating	string	—	e.g. 4 or 4.5
minReviews	integer	—	Drop businesses with fewer reviews
onlyWithPhone	boolean	false	Drop listings with no phone
onlyWithWebsite	boolean	false	Drop listings with no website
categoryKeywords	array	—	Only businesses whose categories mention any of these
maxItems	integer	300	Total cap. 0 = no cap
includeSeen	boolean	true	Off = only businesses not seen in previous runs

---

Output

{
  "id": "yellowpages:473194899",
  "url": "https://www.yellowpages.com/austin-tx/mip/ars-rescue-rooter-473194899",
  "name": "ARS Rescue Rooter",
  "category": "other",
  "categories": ["Plumbers", "Furnaces-Heating", "Plumbing-Drain & Sewer Cleaning"],
  "rating": 2.87,
  "reviewsCount": 15,
  "rank": 1,
  "website": "https://www.ars.com/austin",
  "openingHours": [
    { "day": "Monday", "opens": "09:00", "closes": "17:00" },
    { "day": "Tuesday", "opens": "09:00", "closes": "17:00" }
  ],
  "location": {
    "address": "9411 Atterbury Ln",
    "city": "Austin",
    "state": "TX",
    "postalCode": "78754",
    "country": "US"
  },
  "contact": { "name": "ARS Rescue Rooter", "phones": ["(833) 947-9225"] }
}

Two deliberate choices:

• openingHours is expanded to one entry per day. The site publishes Mo-Fr 09:00-17:00, and you cannot filter "open on Saturday" against that string.
• category is always other. Yellow Pages is a general directory; calling a plumber a restaurant to fill a vertical field would be inventing data. The real classification is in categories, exactly as the site publishes it.

---

Notes and limits

• Yellow Pages only serves US IP addresses and blocks datacenter traffic. This Actor runs behind Apify's unblocking proxy with the country pinned to the US, so you do not have to configure anything — but it is why runs are slower than a plain HTTP scraper.
• The site shows a captcha to traffic that goes too fast. The default rate is deliberately conservative; raising maxRequestsPerMinute a lot is how runs start failing.
• Businesses with no rating are dropped by minRating, because claiming they pass a threshold they never published would misreport the filter you asked for.
• Every listing carries rank, its position on the result page, which is what paid placement looks like from the outside.

---

Pricing

Pay per result. You are charged for businesses that actually land in your dataset — filtered out and duplicate rows cost nothing.