How to search LinkedIn profiles by name, title, company or school

The filters, and the two kinds of name matching

Nothing is required. An empty body is a valid request, which is a good reason to never send one. Name matching comes in two flavours and they behave differently. `search` is fuzzy. Give it a name and it tolerates spelling, middle names and ordering. `first_name` and `last_name` are strict filters. Use `search` when you are looking for a person whose name you heard; use the strict pair when you are matching against a record you already hold and a near-miss would be worse than nothing. Employment splits the same way. `current_company` and `past_company` take a company ID or LinkedIn URL, comma-separated for several. `keywords_company` and `keywords_school` are keyword matches on the name instead, which is what you want when you have a company's name as a string rather than its ID. `follower_of` filters to people following a given profile — the one filter here with no equivalent anywhere else in the catalog, and the way to build an audience list off a founder's following. Geography has a trap. `location` is free text from LinkedIn's autocomplete; `geo_id` is LinkedIn's own identifier and **overrides `location` when both are present**. Resolve the ID with the `geo_id_search` endpoint and pass `geo_id` when you care about the boundary being exact, because free-text locations match approximately and two runs can disagree. `industry_id` works the same way — an ID, not a label. `page` is **1-based**, defaulting to 1.

What you are billed for, on every profile search call

This endpoint is **not** a multi-provider waterfall. Billing is on HTTP 2xx. A thin 2xx still bills. 4xx and 5xx bill nothing, on any endpoint. The miss-free rule on the email finder, phone finder and email verifier does not apply here.

Fields profile search returns

| Field | Type | What it is | |---|---|---| | `elements` | array | The matched profiles. **This array is also the billing unit** — `credits_per_result` is counted off its length. | | `elements[].name` | string | Display name. | | `elements[].linkedinUrl` | string | Profile URL. Feed this straight into [person enrichment](/api/person-enrichment). | | `elements[].position` | string | Headline or current position as LinkedIn renders it, not a normalised job title. | | `elements[].publicIdentifier` | string | The LinkedIn public slug. | | `pagination` | object | **Ships as `{}` in the spec.** Do not write a loop that depends on a key inside it. | Terminate pagination on an empty `elements` array, not on a `pagination` field nobody has confirmed exists. `position` is a headline string, so filtering on it client-side means string matching against whatever the person chose to write about themselves. Treat it as a hint, not a field.

What people use profile search for

**Matching a record you already have to a real profile.** You hold a name and an employer from a form fill or a conference list. Strict `first_name` + `last_name` + `keywords_company` returns one row for a tenth of a credit, and that row gives you the profile URL that keys everything else. **Alumni and past-employer plays.** `past_company` and `school` are the filters behind the two outbound angles that still get replies: people who worked somewhere you have a reference customer, and people who went where your founder went. Neither is expressible in a domain-and-headcount filter set. **Audience lists from a following.** `follower_of` gives you the people following a specific profile. For a founder-led category that is a sharper definition of "interested in this space" than any industry code, and there is no other endpoint here that produces it.

Profile search FAQ

**What am I charged for?** Results returned, measured on the `elements` array, on 2xx responses. There is no per-search fee and no flat call fee. Errors bill nothing. This is not a waterfall endpoint, so the zero-on-miss rule does not apply. **What is the difference between `search` and `first_name`/`last_name`?** `search` matches fuzzily and tolerates variants; `first_name` and `last_name` are strict. Fuzzy finds people whose names you are unsure of and returns more rows, which on per-result pricing also costs more. **Why is my `location` filter being ignored?** Because you also sent `geo_id`, which overrides it. Send one or the other. Resolve `geo_id` through `geo_id_search` rather than guessing an ID. **How is this different from people search?** Different shape, not a different tier. [People search](/api/people-search) filters by company attributes with arrays and 0-based pages; this filters by person attributes with strings and 1-based pages. Pick by whether your question starts with the company or the person. **Can I page through everything that matches?** You can, and on a per-result endpoint that is the expensive way to use it. Narrow the filters until the first page is mostly people you would actually contact. **Do results include emails or phone numbers?** No. You get `name`, `position`, `linkedinUrl` and `publicIdentifier`. Contact details are separate calls: [email finder](/api/email-finder) and [phone finder](/api/phone-finder), both of which bill zero when nobody has the answer. **Can I call this from Claude or Cursor (MCP)?** Yes, the hosted MCP server exposes `profile_search` as a tool with no install. Agents do well with the fuzzy `search` filter, since resolving a half-remembered name is the kind of problem they are given; tell them how many pages they may pull. **Do I need a credit card?** No. 25 free credits on signup, no card.

Frequently asked.

More Search endpoints

See it in a workflow