Find companies similar to a list of seed companies

Required permissions: accounts:read.

Send 1–50 companies by companyId, linkedinUrl, website or name, and a limit (default 100, maximum 500). companyId is an exact company dataset ID (number or numeric string), not a workspace account UUID. When supplied, it takes precedence; a missing ID is reported without trying other fields. Otherwise each company resolves by LinkedIn first, then its marketing website including redirects, then its name. Duplicate matched identities contribute once. No saved accounts are required. Optional hard filters minEmployees, maxEmployees, countries (ISO-2 codes or country names), regions and industries drop candidates outside them, including those with an unknown value for a filtered field unless includeUnknown is set; an employee bound replaces the seed size range.

Returns candidates sorted by similarity to the closest seed, with preliminary high/medium/low bands. Matched seeds and candidates include companyId for reuse in later searches. Check seeds for each matched identity and whether it contributed. Unmatched or unindexed seeds are reported and skipped. If none can contribute, companies is empty. A search failure returns 503. The seed companies themselves are excluded; other existing workspace accounts may appear.

Synchronous, read-only and unmetered. Allow more than 120 seconds for the client timeout. Returns up to limit candidates in one response, with no pagination. Import selected results using POST /v1/accounts/batch.