match_company

match_company resolves companies through OnFire’s matching engine — vector search plus AI matching, so partial, fuzzy, and messy names work.

When to use it

• You have a company name, a website, or a CRM row and need the canonical LinkedIn URL before running ai_prospecting or filtering in ask_onfire.
• You want firmographics (size, employee count) alongside the identity.

Skip it when you already have a company LinkedIn URL. That URL is the canonical OnFire identifier — matching only recovers it when it’s missing. Re-match only when the user explicitly asks you to verify against OnFire.

Inputs

Single-company shortcut — pass the fields directly:

Parameter	Type	Description
name	string	Company name, e.g. "Acme Inc".
website	string	Domain, e.g. "acme.com".
linkedin_url	string	LinkedIn company URL to verify.

Plural forms names, websites, linkedin_urls are accepted too.

Batch form:

Parameter	Type	Required	Description
companies	object[]	Yes	Up to 100 company descriptors — each with names, websites, and/or linkedin_urls.
confirmed	boolean	No	Required true for batches over 50.

At least one of name / website / LinkedIn URL is required per company; multiple signals improve accuracy.

Returns

JSON with total_count, matched_count, and a results array in input order. Each record carries matched, name, website, linkedin_url, linkedin_id, size, found_employee_count, cosine_similarity, match_type, and match_reason.

A batch also returns a dataset handle. To query people at those companies in ask_onfire, filter current_company_url with from_dataset + column instead of re-typing every URL.

Example

{
  "companies": [
    { "names": ["Acme Inc", "Acme Corporation"], "websites": ["acme.com"] },
    { "linkedin_urls": ["linkedin.com/company/globex"] }
  ]
}

Notes & limits

• Batch up to 100 companies per call; over 50 returns needs_confirmation (nothing runs) so the volume can be settled first.
• Results come back in the same order as the input.