Type what you are looking for and where, and get the listings back as rows: the name, the phone, the street, the categories, the rating and review count, the opening hours, the website - and, where the enrichment resolves, contact emails with a deliverability verdict, social profiles and details read off the site itself. Sixty-three columns, one row per listing.
The input is the same two things you would type into the site itself - what you want, and where you want it.
Both fields end up on every row, joined as one query value in the form search, location, so a file that mixes several searches is still separable afterwards.
The directory fields are only the first third of the row. The rest is enrichment: emails with a deliverability verdict, social profiles, and details read off the business's own website.
Addresses arrive alongside a status rather than on their own, so a list can be filtered down to what is worth sending to before anyone loads it into a sequencer.
Opening times arrive as Mo-Fr 09:00-18:00, Sa 09:00-15:00 rather than prose, which is a form you can parse instead of a sentence you have to interpret.
This export has real data in it, so this page can be specific about formats. It also has one property that will quietly corrupt a pipeline, and it is worth two paragraphs before the table.
The column order is not stable between runs. Every run we have carries the same sixty-three names, but they do not always arrive in the same sequence - in one run hours sat second-to-last rather than twentieth, and the columns after it shifted by one. Nothing was added, removed or renamed; only the order moved.
That is harmless if you read by column name and destructive if you read by position. A loader that takes column 20 as amenities because it was amenities last week will fill that field with opening hours and never raise an error. Map the header row by name on every import - and if your tooling makes that awkward, this is the one thing worth fixing before the first real run rather than after it.
Taken from the header row and the JSON keys of real runs, which agree. Formats below are measured from populated cells - where a column has a fixed shape, it is given; where it does not, it is not invented.
(212) 473-4444 - parentheses, a space, then a hyphen.categories joined with a comma and a space, so the two are the same fact twice - take whichever suits you and ignore the other.reviews_count always arrive together - a listing has both or neither.open now, closed now, opening soon, open 24 hours. Read it against that row's scraped_at, not against the run; see the note below.email_1 to email_3 and emails, which are the columns to read for contact addresses.Mo-Fr 09:00-18:00, Sa 09:00-15:00 - two-letter day codes, ranges joined by commas.emails, each two comma-joined parts: a deliverability verdict, then a detail. The detail is not a fixed vocabulary - it can name the kind of mailbox or echo a domain - so switch on the part before the comma and treat the rest as free text.<title> of the business's site.phone, so de-duplicate against that column before dialling anything.phones_extra, positionally aligned and always the same length - but an entry can be null where the type could not be determined, so index into it defensively.phone.contact_name.2026-07-06 09:16:41.954000. This varies within a single run - rows in one file were not all read at the same instant, which is what makes open_status a per-row fact.search, location, repeated on every row that came from them.email_2 without this one.Read this export by column name, never by position. Every run carries the same sixty-three names, but the order is not fixed - in one of ours hours arrived second-to-last instead of twentieth, and everything after it shifted by one. Nothing was renamed or dropped; a positional loader would simply have written opening hours into the amenities field and reported success. The second thing to know is that scraped_at varies within a single file, because the rows were not all captured at the same instant - so open_status describes each row's own moment and is not a property of the run. And three smaller shapes worth handling before they surprise you: category is just categories joined by a comma, so it is the same fact twice; amenities can pack two values into one string separated by a newline; and phones_extra_types lines up with phones_extra index for index but can hold a null where the type was not determined.
Four controls, and they are the whole input: where to search, what to look for, where to look, and how much to take. The sort and filter buttons beside the results are view controls - they shape what you are looking at, not what the run fetches.
A few examples of how teams use directory listings to answer a question they actually have.
A category and a city produce the listings; the enrichment columns produce the way to reach them. Because each address arrives with a deliverability verdict, the list can be filtered before it ever reaches a sending tool rather than after the bounces come back.
Rating, review count and years in business sit on the same row, which is enough to tell a crowded market from a thin one and a new entrant from a fixture - without opening a single listing by hand.
The website columns say what built each site and whether it carries a Facebook pixel or a Google tag. For anyone selling to a platform's users, that turns a directory into a qualified list.
No subscription, no minimum, no recurring bill. Your first 500 rows are on us - after that you pay as you go, at the same flat rate as every other scraper here.
On every new account, one-time. No credit card. The record limit and every export format are included.
About $2 per 1,000 listings, enrichment columns included rather than billed as an extra step. The estimator shows the row count and credit cost before a run starts.
Volume pricing, dedicated workers and an SLA for ongoing monitoring or very large historical pulls. Tell us your numbers and we will put together a quote.
LIVESCRAPER10 at checkout.The same category-and-city job against the other big local index. Running both is how you find the businesses one source lists and the other has never heard of.
For the rows where the built-in enrichment came back without an address. Feed it the websites from this export and work the gap rather than accepting it.
Short answer: yes for the listing content - but this export carries contact data and a deliverability verdict, and that is the part that deserves your attention.
A Yellow Pages listing exists to be found. The name, phone, address, categories, hours and rating are shown to anyone who runs the search, signed in or not, and collecting publicly listed business information for market research is long-settled practice. Nothing here touches a login, an account or a paywall.
The enrichment is where care is owed. This export does not stop at the directory: it adds email addresses, the people attached to them, extra phone numbers with line types, and social profiles. A business address at a company domain is a business contact in most readings - but a named individual's address is personal data in the EU and the UK regardless of where it was published, and emails_persons exists precisely to name individuals. If you are doing outreach, the lawful basis and the opt-out are yours to get right, and the deliverability verdict in emails_status tells you whether a message will arrive, not whether you are allowed to send it. The two are unrelated and it is worth keeping them separate in your own head.
If your question is about market structure rather than contact - how many, how established, how well rated - drop the email, phone and person columns at ingest and most of this stops applying to you. We run no third-party trackers on the data layer, and your exports self-delete after 30 days. Yellow Pages' own terms restrict automated access, so check them before you scale.
The questions we hear most. Something else? Talk to us - humans write the answers, not bots.
500 one-time free rows on every new account - no expiry date. After that $0.002 per row, pay as you go - no card on file until you say so.
Livescraper's Yellow Pages Search Scraper turns a category and a location into a lead table. You pick a region, type what you are looking for and where, optionally cap the number of records, and download the results as a clean CSV, Excel or JSON file. One row comes back per listing, and both of your inputs are carried on every row as a single query value so a file that mixes several searches stays separable.
Each row carries sixty-three columns. The directory side is the business name, phone, street, city, state and postcode, the categories both as an array and as a joined string, the rating and review count, the opening hours in a compact machine-readable form, a description snippet, an image and a link back to the listing. The enrichment side adds email addresses with a deliverability verdict, the people attached to them, further phone numbers with their line types, social profiles, and details read off the business's own website - its title, description, keywords, what built it, and whether it carries a Facebook pixel or a Google tag.
Sales teams use it to build a contactable list in a single pass, filtering on the deliverability verdict before anything reaches a sending tool. Researchers put rating, review count and years in business side by side to tell a crowded local market from a thin one. Go-to-market teams filter on the website columns to find businesses running a particular platform.
One property of this export deserves attention before you build on it: the column order is not stable between runs. Every run carries the same sixty-three names, but in one of ours the hours column arrived second-to-last rather than twentieth and the columns after it shifted by one. Nothing was renamed or dropped, which is exactly why a loader that reads by position mis-maps silently and reports success - map the header row by name on every import. It is also worth knowing that scraped_at varies within a single file, so the open_status field describes each row's own moment rather than the run as a whole. Start free: your first 500 rows cost nothing and need no credit card, and after that it is a flat $0.002 per row.