LinkedIn Company Headcount & Team Directory Scraper

LinkedIn Company Headcount & Team Directory Scraper

Get a company's LinkedIn-reported headcount and firmographics without logging in — employee count, size band, followers, industry, headquarters, type, website, address, about text, specialties, and logo. Add your own LinkedIn session cookie only when you also want the optional employee roster.

What this actor does

• Cookieless headcount first: one recordType: company row per input with LinkedIn's reported employee count, size band, followers, industry, HQ, type, website, address, specialties, and logo.
• Domain-aware company resolution: accepts LinkedIn URLs, company domains, and names; mixed inputs are allowed. A URL is trusted, a domain is confirmed against the page's website (including rebranded hosts such as notion.so → notion.com), and a name must match the page exactly with a matching website or enough follower authority.
• Honest ambiguity handling: bare-name inputs such as Notion, Linear, or Cal can collide with a different same-name LinkedIn page. Those return an explicit error row instead of a silent wrong-company match; domain forms such as cal.com resolve normally.
• Optional cookie roster: with a valid cookie, fetches up to 2,500 visible employees per company in 10-row pages, with title, location, seniority, function, connection degree, language, and resume-from-page filters.
• Explicit coverage and confidence: every company row carries headcountConfidence, employeesReported, employeesReturned, and a coverage block. Error rows have an errorCode, so a missing roster never looks like a complete one.
• Empty fields are omitted: missing values are left out instead of being returned as null.

Output per company (recordType: "company")

• input — the value you supplied
• companyName, companySlug, companyUrl
• linkedinCompanyId — numeric LinkedIn organization ID
• resolutionMethod — provided_url, domain_slug, or name_variant
• resolutionConfidence, confidence — high for a validated page
• resolutionCandidates — alternate slugs tried, when any
• linkedinEmployeeCount — LinkedIn-reported employee count
• linkedinEmployeeCountSource — public_company_page
• companySizeRange — LinkedIn's band, e.g. 5,001-10,000 employees
• headcountConfidence — high when the count is inside the band or the band is open-ended; medium up to 2x above; low above 2x
• headcount — { reported, sizeBand, asOf, source, audited: false }; the count is never described as audited
• followerCount, industry, headquarters, companyType
• website, primaryAddress, address
• about, tagline, specialties[], locations[], logoUrl
• employeesReported — LinkedIn's company count
• employeesReturned, employeesCoveragePct
• rosterMode — cookieless or cookie
• rosterAvailable, rosterReportedTotal, rosterRowsReturned, rosterCoveragePct, rosterConfidence, rosterReason
• coverage — { employeesReturned, employeesReported, rosterReportedTotal, rosterCoveragePct, rosterAvailable, rosterConfidence }
• filtersApplied — the roster filters that were sent
• sourceMode — cookieless_public_page
• warnings[] — human-readable caveats
• recordType, scrapedAt

Output per employee (recordType: "employee")

• companySlug, companyName, currentCompany
• name, firstName, lastName, publicIdentifier
• headline, currentTitle, currentDepartment, department
• seniority, senioritySource — derived from the title when LinkedIn does not expose a seniority field
• location, country
• profileUrl
• profilePicture, profileImageUrl, profilePictureUrl
• tenure, tenureYears, tenureSource — only when LinkedIn exposes a duration
• connectionDegree
• industryName, currentCompanyLogo
• openToWork, premium, verifiedStatus — only when the badge is present
• email, emailSource — only if an email is explicitly visible; usually absent
• source — cookie-roster
• sourceMode — cookie_voyager
• companyFilterApplied — true; rows are only returned when the company filter was applied
• confidence — high with a headline, medium otherwise
• recordType, scrapedAt

Output per error (recordType: "error")

• input, companySlug, companyName (when known)
• errorCode — ambiguous_company_name, company_not_found, company_unresolved, rate_limited (the public company page returned 999/429), cookie_missing, cookie_invalid, company_id_unresolved, roster_rate_limited, roster_request_budget_reached, or roster_unavailable
• error, message, help
• retryable
• resolutionCandidates, rejectedCandidates — for resolution failures
• recordType, scrapedAt

Input

Provide at least one of companyUrls, companyDomains, or companyNames. You can mix them in one run.

Example: cookieless headcount for mixed inputs

{
  "companyUrls": ["https://www.linkedin.com/company/stripe/"],
  "companyDomains": ["notion.so", "linear.app"],
  "includeEmployees": false,
  "maxItemsTotal": 10
}

Example: batch headcount by name

{
  "companyNames": ["Microsoft", "HubSpot", "Figma"],
  "includeEmployees": false,
  "maxItemsTotal": 20
}

Example: roster for one company with a cookie

{
  "companyDomains": ["raycast.com"],
  "includeEmployees": true,
  "maxEmployees": 50,
  "titleKeywords": "Engineer",
  "seniorityLevels": ["Mid-Senior", "Director"],
  "maxItemsTotal": 100,
  "cookies": "YOUR_LI_AT_COOKIE"
}

Example: resume a capped roster

{
  "companyUrls": ["https://www.linkedin.com/company/microsoft/"],
  "includeEmployees": true,
  "maxEmployees": 100,
  "startPage": 26,
  "maxItemsTotal": 200,
  "cookies": "YOUR_LI_AT_COOKIE"
}

Use cases

• Account research — capture headcount, size band, HQ, industry, and website for a list of target companies without a LinkedIn login.
• Market mapping — compare headcount and follower counts across a segment of public and private companies.
• Recruiting and sourcing — pair cookieless firmographics with an optional roster when you have a session that can see the people tab.
• Sales operations — enrich CRM accounts with LinkedIn-reported size and industry fields.
• Competitive tracking — periodically snapshot headcount, followers, and specialties for competitor pages.
• Data quality checks — use headcountConfidence to flag companies whose reported count and size band disagree.

Limitations and caps

• Headcount is LinkedIn-reported, not audited. The public count can differ from LinkedIn's own size band; both fields are returned and headcountConfidence labels the gap. It is not a payroll, HRIS, or verified workforce count.
• The roster is optional and cookie-gated. Without a valid cookie only the company row is returned, with an explicit warning and no employee rows.
• LinkedIn caps a roster query at 2,500 visible results. Pages contain 10 rows, so a query can need up to 250 requests. The actor budgets about 250 roster requests per run to stay inside LinkedIn's daily limit; when the budget is used, the remaining companies get an explicit error row and a later run can resume with startPage. Large companies are samples rather than complete workforces.
• maxEmployees is the per-company cap. If LinkedIn reports more employees than the cap, the row set is a sample; rosterReportedTotal and rosterCoveragePct make the shortfall visible.
• Filters are best-effort. Title, seniority, function, connection degree, and language are sent to LinkedIn; location is matched against the returned location text. Missing fields are not used to exclude a row.
• Email is not harvested. It is returned only when an email is explicitly visible on the result; expect none. Tenure is returned only when LinkedIn exposes a duration.
• Past employees are not supported. The actor returns current employees when a roster is requested.
• Name-only resolution is restricted. Ambiguous bare names (Notion, Linear, Cal, and similar) return an error row. Pass the LinkedIn URL or the company domain instead (for example cal.com for Cal.com).
• Public pages can rate-limit. If LinkedIn throttles the company page itself, the actor emits a retryable rate_limited error row instead of a company row. If it throttles a roster request, the company row is still returned with a warning. Retry later or narrow the input list.

FAQ

Do I need a LinkedIn cookie for headcount? No. Company name, LinkedIn-reported employee count, size band, followers, industry, HQ, type, website, address, about, specialties, and logo come from the public company page.

When do I need a cookie? Only for the employee roster. Set includeEmployees to true and add a valid li_at value or full cookie export. The cookie is never required for the company row.

Why did I get an error row instead of employees? The run tells you which case applies: cookie_missing (no cookie), cookie_invalid (expired or rejected session), company_id_unresolved (the roster filter could not be mapped), roster_rate_limited (LinkedIn returned 999/429), or roster_request_budget_reached (the ~250-request run budget was used). Company headcount is still returned whenever the page resolves.

Why does the employee count disagree with the size band? LinkedIn shows two independent fields. A company can report 7,000 employees while its self-selected band says 501–1,000. The actor publishes both and labels the count high, medium, or low confidence rather than presenting either as audited.

Why is the roster smaller than the company's headcount? LinkedIn caps a single query at 2,500 visible results and serves 10 per page; the daily request limit is about 250. For a large employer, the roster is a sample. Raise maxEmployees, split the request with startPage, or narrow the filters.

Why did my company name resolve to a different company? Some names are shared by several LinkedIn pages. The actor refuses the low-confidence name match and asks for a URL or domain. For example, notion.so resolves to the right Notion page, while the name alone is treated as ambiguous.

Can I get email addresses? Only if LinkedIn explicitly shows an email in the result. The actor does not guess or construct emails, so most records have no email field.

Can I get tenure? Only when LinkedIn exposes a duration for the person. Otherwise tenure and tenureYears are omitted instead of estimated.

Does the actor return past employees? No. The roster is current employees only.

Is this affiliated with LinkedIn? No. This is an independent, third-party actor. It uses publicly visible company pages and, when you supply one, your own LinkedIn session for the roster. You are responsible for how you use the data.