LinkedIn Companies Scraper

Extract the data LinkedIn exposes publicly on supplied company profile URLs. This actor is for public company-profile pages only: it does not request a login, cookies, employee search, hidden tabs, or authenticated/private endpoints.

Input

{
  "companyUrls": [
    "https://www.linkedin.com/company/apify/",
    "https://www.linkedin.com/company/microsoft/"
  ],
  "maxCompanies": 10
}

Use public URLs matching https://www.linkedin.com/company/<company-slug>/. The actor does not resolve a company name or website domain into a LinkedIn URL, because that would require a separate search data source and would make the result less predictable.

Field	Default	Notes
companyUrls	[]	Up to 100 public /company/<slug>/ URLs per run. Empty input returns one labeled, uncharged sample row.
maxCompanies	10	1–100. Caps how many valid URLs are actually requested, so you can paste a long list and still bound the spend.
requestTimeoutSecs	20	5–60 seconds for a single company page request.

Output

Each real row contains only fields found in a public LinkedIn page's Organization JSON-LD or public metadata:

• name, description, companyUrl, and companySlug
• followerCount when LinkedIn exposes it in public metadata
• employeeCount when LinkedIn exposes a public JSON-LD value
• headquarters and structured address when exposed
• website and logoUrl when exposed
• transport (direct_http) and retrievedAt

Missing public fields are null; the actor does not infer industry, funding, employee lists, contacts, revenue, or data behind a sign-in wall.

Diagnostics instead of silent gaps

Every URL that does not produce a company row produces a labeled diagnostic row instead: BAD_INPUT (not a /company/ URL), BLOCKED (LinkedIn's 999/403 or a security-verification interstitial), RATE_LIMITED, NOT_FOUND, or NO_RESULTS when the page loads but carries no matching public Organization JSON-LD. You can always reconcile input count against row count.

Proxy policy

The actor uses direct HTTP only. It never starts a residential proxy, so a blocked page returns an uncharged diagnostic rather than quietly spending proxy credit.

Empty input

Empty input produces exactly one _sample: true sample row. It is clearly labeled and never charged for a company, so the actor can be evaluated without a LinkedIn request or proxy use.

Pricing

$0.90 per 1,000 companies ($0.0009 each), plus $0.001 each time a run starts. Flat rate — no volume tiers, no plan gates — and only rows that parse as genuine public company profiles trigger the company charge event.

Invalid input, blocked pages, not-found and parse failures, duplicates, samples and diagnostics never incur the per-company fee; a run that returns nothing costs only the $0.001 start fee. Rows are written to the dataset before the charge call, so you can always audit what you paid for. Because companyUrls takes 100 URLs in one run, a full 100-company batch costs $0.09 plus a single $0.001 start.