generated: '2026-06-20'
method: searched
source:
- https://learn.microsoft.com/en-us/linkedin/shared/api-guide/concepts/pagination
- https://learn.microsoft.com/en-us/linkedin/marketing/versioning
- https://learn.microsoft.com/en-us/linkedin/shared/api-guide/concepts/error-handling
- openapi/*.yml
protocol:
  name: Rest.li
  protocol_version_header: X-Restli-Protocol-Version
  protocol_version: 2.0.0
  note: >-
    LinkedIn APIs are built on the open-source Rest.li framework. Identifiers are
    expressed as URNs (urn:li:<domain>:<id>). Requests must pin the Rest.li
    protocol with X-Restli-Protocol-Version: 2.0.0.
authentication:
  style: OAuth 2.0 Bearer
  header: 'Authorization: Bearer {token}'
  flows: [authorizationCode, clientCredentials]
  oidc: Sign In with LinkedIn uses OpenID Connect (openid, profile, email scopes)
  see: authentication/linkedin-authentication.yml
versioning:
  header: Linkedin-Version
  format: YYYYMM
  required: true
  see: lifecycle/linkedin-lifecycle.yml
pagination:
  style: index-based (offset)
  params:
  - name: start
    description: Index of the first item to return.
    default: 0
  - name: count
    description: Number of items per page (fewer may be returned).
    default: 10
  response_object: paging
  response_fields: [start, count, total, links]
  results_field: elements
  end_of_dataset: fewer elements returned than the requested count
  cursor_note: >-
    Some newer/finder APIs use metadata cursor tokens; where present the response
    carries a paging.links[] with rel next and a start/count derived from the cursor.
field_projection:
  mechanism: Rest.li field projections
  param: 'fields (e.g. ?fields=id,firstName,lastName) and decoration via projection syntax'
  note: >-
    LinkedIn supports Rest.li decoration/field projection to shape response
    payloads and expand referenced URN entities inline.
batching:
  mechanism: Rest.li batch operations
  patterns: [BATCH_GET (ids=List(...)), batch create, batch update, batch delete]
  note: Multiple entities addressed in one call via the ids query parameter with Rest.li encoding.
request_tracing:
  note: >-
    LinkedIn returns an x-li-uuid / x-li-fabric response header on many endpoints
    that uniquely identifies the request for support escalation.
error_envelope:
  content_type: application/json
  fields: [status, message, serviceErrorCode]
  format: LinkedIn/Rest.li error (not RFC 9457 problem+json)
  see: errors/linkedin-problem-types.yml
rate_limiting:
  model: per-application and per-member daily throttles
  signal: HTTP 429 (Too Many Requests)
  docs: https://learn.microsoft.com/en-us/linkedin/shared/api-guide/concepts/rate-limits
  see: rate-limits/linkedin-rate-limits.yml
idempotency:
  header: null
  note: >-
    LinkedIn does not document a general-purpose idempotency-key header; create
    conflicts surface as 409 (e.g. uniqueForeignId already exists). Some
    resources accept a client-supplied uniqueForeignId to dedupe.