Bark.com Scraper — company profiles, ratings & reviews by category and city

Turn any Bark.com category + US city into a structured list of local service companies, complete with their aggregate rating, review count and individual customer reviews.

Bark is one of the largest home- and professional-services marketplaces. Its public category × city pages list the companies competing in a market; each company profile carries schema.org microdata with the business identity and its full public review history. This actor reads that structured data directly — no browser, no private API, no guessing at selectors.

Use it for competitive/review intelligence: track how rivals are rated in the cities you serve, mine review text for service gaps, size a market before you enter it, or build a ranked shortlist of providers in a category.

What this actor does NOT return: phone, email and postal address. Bark deliberately keeps those behind a JavaScript "Reveal" / lead-form flow, so they are not present in the page at all. This is a review-intelligence actor, not a contact-enrichment actor. Nothing is silently null — those fields are simply not offered.

What you get

• Business identity: name, Bark profile URL, stable barkCompanyId, description, logo
• Services the company serves (categoriesServed) and the category/city it was found under
• Reputation: aggregate rating, total review count
• Up to ~50 individual reviews per company: author, star rating, date and full review text

Input

Field	Type	Default	Description
mode	string	search	search = crawl category × city listings. startUrls = scrape the exact Bark URLs you supply.
category	string	accountants	Bark category slug, e.g. accountants, plumbers, photographers, personal-trainers.
locations	array of strings	–	US cities in City, ST form, e.g. ["Austin, TX", "New York City, NY"]. Both the 2-letter code and the full state name are accepted.
startUrls	array	–	Used when mode is startUrls. Accepts Bark listing URLs (/en/us/<category>/<state>/<city>/) and company profile URLs (/en/us/company/<slug>/<code>/).
includeReviews	boolean	true	Emit the individual reviews array per company.
maxReviewsPerCompany	integer	20	Cap on reviews stored per company.
maxItems	integer	100	Hard cap on companies per run. 0 = unlimited.
proxyConfiguration	object	Apify RESIDENTIAL, country US	Required. Bark serves a CloudFront 403 to datacenter and direct requests; residential US returns clean 200s.

Run it with no input at all and the actor performs a small documented sample run — accountants in New York City, NY, capped at 10 companies — so you always get real example rows.

Output

One row per company:

Field	Description
barkCompanyId	<slug>/<code> — stable dedupe key
businessName	Company name
profileUrl	Bark profile URL
description	Company's own description
categoriesServed	Services the company lists (schema.org knowsAbout)
aggregateRating, reviewCount	Aggregate reputation
reviews[], reviewsShown	Individual reviews (author, rating, date, text) and how many were stored
logoUrl	Company logo
scrapedAt	ISO timestamp of the fetch
searchCategory, searchCity, searchState	The query this row was found under
sourceUrl	The listing page the company was discovered on

Duplicates are collapsed run-wide on barkCompanyId, so the same company appearing in several cities or categories is only emitted — and only charged — once.

Reliability

• The run ends successfully with a status message if Bark blocks the transport or a query has no results; it never fails a run over a transient block. Re-running picks fresh residential IPs.
• A wall-clock budget stops the crawl before the platform's run timeout and emits everything already collected, rather than being killed mid-run.

Pricing

Pay-per-event: you are charged only for company rows actually delivered to the dataset (company-scraped). Blocked fetches, duplicates and pages that yield nothing cost you nothing. See the Pricing tab for the current per-event price.

Legal

Only ordinary public page loads are used. You are responsible for complying with Bark's Terms of Service and with applicable data-protection law — review text and reviewer display names are personal data.