Skip to content
Parsium Docs

LinkedIn Prebuilt APIs

All LinkedIn prebuilts cost 1 credit and accept GET query parameters or POST JSON.

Profile

Section titled “Profile”

/v1/linkedin-profile accepts a profile url or username. LinkedIn guest pages can mask job titles and schools with asterisks. Parseium returns those values as null and adds a masked_content warning listing affected field paths.

LinkedIn HTTP 999/429 responses and auth walls are retried with backoff and fresh proxy connections. If retries are exhausted, the endpoint returns retryable rate_limited and consumes no credit.

Company

Section titled “Company”

/v1/linkedin-company accepts a company url or slug, plus:

  • posts_limit (1–100, default 10)
  • posts_offset
  • posts_since (RFC3339 or YYYY-MM-DD)

The response includes metadata, normalized domain, guest-visible posts, employee profile URLs, and a similar-company discovery graph. LinkedIn’s guest page currently exposes at most 10 posts; posts_page.guest_page_cap makes this explicit.

Post

Section titled “Post”

/v1/linkedin-post accepts common feed/update and /posts/ URLs, Pulse URLs/slugs, and bare activity IDs through url or slug. Feed input is normalized for reliable fetching; canonical_url contains LinkedIn’s canonical result URL.

The endpoint returns full guest-visible text, author, images, reaction count, and comments. LinkedIn does not expose complete commenter/reactor identities to guest sessions. Do not infer identities that are absent; commenter profile URLs are usually unavailable.

Structured errors

Section titled “Structured errors”

Use code, not message, when branching: unsupported_type (422), rate_limited (429), not_found (404), and fetch_failed (502). Respect retry_after_seconds for retryable errors. Failed calls do not consume credits.