All notable changes to LinkedIn Company Scraper are recorded here. The newest release is first.
New input and more reliable name lookups. Existing input fields, output fields and prices are unchanged.
- New Website domains to match (
domains) input: give company website domains (stripe.com) and get each company's LinkedIn page. The Actor follows the LinkedIn link on the company's own homepage, then tries linkedin.com/company/<name>, then LinkedIn's company search, and keeps a page only when it is verified. New columns inputDomain, matchedBy and matchConfidence; a domain with no verified page gets a free not_found row.
- Typed company names and numeric ids now find companies whose open jobs are all outside the country of the proxy IP (Grab, Reliance Industries, Safaricom): the job search behind the lookup is worldwide.
- When a guessed page address does not exist, the next guess is tried instead of ending with
not_found (Dangote Group now resolves to Dangote Industries Group).
- A typed name whose company was also given as a URL, or two names for the same company, now each get their own row (before, the second one got no row at all).
Reliability fixes. Output fields, input fields and prices are unchanged.
- A run that Apify restarts partway (a server migration) keeps counting what it already saved. Before, such a run could end as FAILED with all its rows in the dataset, write the rows for unusable input entries twice, and start counting
maxItems again from zero.
- Company addresses that LinkedIn redirects to a showcase page (for example
/company/gojek/ and /company/expedia/) now return the company's row instead of an error row. Showcase URLs (linkedin.com/showcase/...) and Sales Navigator company links (linkedin.com/sales/company/<id>) are accepted as input.
- When LinkedIn answers HTTP 999 (its "request denied" code), the request is retried from another IP after a short pause, the same as an HTTP 429.
- Company ids sent as JSON numbers (
[1035]) are looked up instead of being skipped silently. A single text value instead of a list no longer stops the run.
- If the proxy chosen in the input cannot be used when the run starts, the run falls back to Apify Proxy's default group (or no proxy) with a warning, instead of failing before it fetches anything.
First public release.
- Reads LinkedIn's public company pages as a logged-out visitor: no account, no cookies, no login, no browser.
- Companies by URL (any country site or sub-page), slug, numeric id, or typed company name (LinkedIn's public company search).
- Company facts only: name, website, industry, size range, headquarters and address, founded, type, specialties, followers, logo, cover image, office locations, similar and affiliated pages, open jobs. No employees, people or posts.
- Every row carries
status, so a missing company (not_found) and a failure (error) are told apart from real data. Only ok rows are charged.
- Callable with no arguments by an AI agent: the default input returns 6 companies in seconds.
- Typed company names also work for companies with no open jobs on LinkedIn (for example Safaricom, Toyota, Jumia): the Actor tries the page address that matches the name and keeps only a page with the right company id.