Evidence-Verified LinkedIn Company Resolver

Resolve company domains or website URLs to verified LinkedIn company URLs using first-party website evidence.

This Actor is built for workflows where a plausible-looking LinkedIn URL is not good enough.

Instead of guessing from search-engine results, it looks for evidence published on the company's own public web properties and returns a LinkedIn company URL only when the evidence is strong enough.

Precision over recall: if the evidence is weak, conflicting, or ambiguous, the Actor returns resolved: false instead of guessing.

---

Why use this Actor?

• First-party website evidence only
• No LinkedIn scraping
• No LinkedIn login
• No cookies
• No Google, Bing, Exa, SerpAPI, or other external search API
• No paid external API
• No LLM-generated guesses
• Evidence URL returned with successful matches
• Verification method included
• Confidence score included
• Explicit failure / non-resolution reasons
• One output row for every submitted input
• Designed for batch enrichment and automated workflows

---

Common use cases

• B2B lead enrichment
• CRM data enrichment
• Domain-to-LinkedIn matching
• Company identity resolution
• Account-based marketing (ABM)
• Sales intelligence pipelines
• Lead database cleanup
• AI agent workflows
• Make and n8n automations
• API-based enrichment pipelines

---

How verification works

The Actor searches for LinkedIn company URLs in evidence published on the company's own public web properties.

Evidence may include:

• Organization.sameAs in JSON-LD
• LinkedIn company links published on first-party pages
• LinkedIn URLs embedded in first-party HTML or application payloads
• First-party About, Company, Contact, Press, or similar pages
• Same-company subdomains
• Legitimate public HTTPS domain migrations
• First-party sitemap-discovered pages

The Actor may use limited first-party site discovery to locate relevant pages, but it does not use external search engines to manufacture a result.

A LinkedIn URL is returned only when the evidence meets the resolver's verification requirements.

---

Input

Basic input:

{
  "domains": [
    "stripe.com",
    "notion.so",
    "shopify.com"
  ]
}

Website URLs are also accepted:

{
  "domains": [
    "https://stripe.com",
    "https://www.notion.so",
    "https://www.shopify.com"
  ]
}

Optional controls

• maxPagesPerDomain — default 5
• maxConcurrency — default 10
• requestTimeoutSecs — default 15
• domainTimeoutSecs — default 90
• includeDiagnostics — default false

For most users, the default settings are recommended.

---

Output

Each submitted input produces one dataset row.

Resolved example

{
  "input": "stripe.com",
  "domain": "stripe.com",
  "canonicalDomain": "stripe.com",
  "resolved": true,
  "linkedinCompanyUrl": "https://www.linkedin.com/company/stripe/",
  "linkedinSlug": "stripe",
  "evidenceUrl": "https://stripe.com/",
  "verificationMethod": "json_ld_organization_same_as",
  "confidence": 1.0,
  "reason": null,
  "checkedAt": "2026-10-05T00:00:00Z"
}

Canonical-domain example

A submitted domain may legitimately redirect to a newer official domain.

For example:

{
  "input": "notion.so",
  "domain": "notion.so",
  "canonicalDomain": "notion.com",
  "resolved": true,
  "linkedinCompanyUrl": "https://www.linkedin.com/company/notionhq/",
  "linkedinSlug": "notionhq",
  "evidenceUrl": "https://www.notion.com/",
  "verificationMethod": "embedded_first_party_url",
  "confidence": 0.99,
  "reason": null,
  "checkedAt": "2026-10-05T00:00:00Z"
}

Unresolved example

{
  "input": "example.com",
  "domain": "example.com",
  "canonicalDomain": "example.com",
  "resolved": false,
  "linkedinCompanyUrl": null,
  "linkedinSlug": null,
  "evidenceUrl": null,
  "verificationMethod": null,
  "confidence": 0.0,
  "reason": "no_first_party_evidence",
  "checkedAt": "2026-10-05T00:00:00Z"
}

---

Output fields

Field	Description
input	Original submitted domain or URL
domain	Normalized submitted domain
canonicalDomain	Final verified first-party domain when applicable
resolved	Whether a verified LinkedIn company URL was found
linkedinCompanyUrl	Verified LinkedIn company URL
linkedinSlug	LinkedIn company slug
evidenceUrl	First-party page containing supporting evidence
verificationMethod	Method used to verify the match
confidence	Resolver confidence score
reason	Explanation when resolution fails
checkedAt	Timestamp of the resolution attempt

Optional diagnostic fields may also be returned when includeDiagnostics is enabled.

---

Pricing

You pay only for successfully resolved companies.

• $4.00 per 1,000 resolved companies
• $0.004 per successfully resolved company
• Unresolved results do not trigger the resolved-company charge
• Platform usage is included
• A small Actor start fee of $0.00005 applies per run

The resolved-company event is triggered only when:

resolved: true

Not-found, timeout, blocked, ambiguous, and other unresolved results are returned to the dataset without triggering the resolved-company charge.

---

Precision-first behavior

This Actor is intentionally designed to minimize false positives.

A LinkedIn link found somewhere on a website is not automatically assumed to belong to that company. It may belong to a partner, employee, customer, investor, or another organization.

The resolver therefore requires sufficiently strong first-party evidence before returning a result.

This means the Actor may resolve fewer domains than search-based tools that aggressively guess.

If your workflow values verified matches over maximum match rate, this Actor is designed for that use case.

---

Possible unresolved reasons

Typical reasons include:

• no_first_party_evidence
• weak_or_uncorroborated_evidence
• ambiguous_multiple_company_urls
• http_403
• http_429
• cross_site_redirect_blocked
• domain_timeout
• non_html_content
• internal_error
• charge_limit_reached

These reasons make it easier to decide whether an unresolved company should be retried, reviewed manually, or handled by another step in your workflow.

---

Charge-limit behavior

The Actor respects Apify's pay-per-event charge limits.

If the user's maximum charge limit is reached:

• pending resolution work is stopped
• already completed results remain available
• unprocessed inputs are still returned
• those rows use:

reason: "charge_limit_reached"

This prevents silent data loss in bulk runs.

---

Security

The Actor:

• accepts only public web targets
• blocks private and non-public network targets
• blocks loopback, link-local, multicast, reserved, and unspecified IP ranges
• rejects unsafe cross-site redirects
• limits response body size
• checks only HTTP/HTTPS resources
• does not accept credentials embedded in input URLs

---

What this Actor does not do

This Actor deliberately does not:

• scrape LinkedIn pages
• log in to LinkedIn
• use LinkedIn cookies
• scrape employee profiles
• return contacts or personal data
• scrape LinkedIn posts
• use an LLM to guess matches
• use Google, Bing, Exa, SerpAPI, or another paid search API

The product is intentionally narrow: company domain → evidence-verified LinkedIn company URL.

---

Reliability

The Actor is designed so that one difficult website does not fail the whole batch.

Slow, blocked, malformed, or otherwise problematic domains return structured unresolved rows while the rest of the batch continues processing.

One output row is returned for every submitted input.

---

Validation

The current production version was validated on a 100-company test set.

Results:

• 100 submitted inputs
• 100 output rows
• 62 verified resolutions
• 62% resolution rate
• No obvious false positive was identified during sample verification
• Unresolved cases were returned with explicit reasons

The resolver intentionally prioritizes precision over maximizing the resolution rate.

---

Production status

Current version: v0.1.7

The resolver logic is considered production-ready.

Future changes will primarily address meaningful reliability issues discovered through real-world usage rather than increasing resolution at the expense of precision.