generated: '2026-08-14'
method: searched
source: https://docs.fullcontact.com/docs/rate-limiting
description: >-
  FullContact publishes both its default ceiling and the exact response headers
  it returns, which is the part an agent needs. The signalling is unusual:
  rather than reject immediately, FullContact DELAYS a request up to 1000ms to
  keep it inside the configured limit and reports the delay it applied in
  X-FullContact-RateDelay. A 429 is only returned when the required delay would
  exceed 1000ms, and that 429 is itself delayed by the maximum latency so the
  caller can retry immediately and safely.

limit_count: 1

limits:
  - scope: per-account (API key)
    window: 60s
    limit: 6000
    unit: requests per minute
    burst: null
    note: >-
      "Our default rate limit is 6,000 QPM but should you need a higher rate
      limit contact our Sales team to get you setup." Rate limits exist
      independently of the key's monthly usage allowance.
    upgrade: https://www.fullcontact.com/contact-us/

headers:
  - name: X-FullContact-RateDelay
    meaning: >-
      Milliseconds of latency FullContact added to this request in order to
      keep it inside the configured rate limit.
    example: '5'
    authoritative: true
  - name: X-Rate-Limit
    meaning: Static rate limit ceiling. Legacy header retained for backwards compatibility on older client libraries.
    example: '600'
    authoritative: false
  - name: X-Rate-Limit-Remaining
    meaning: Static number of requests left in the 60-second window. Legacy, static value.
    example: '10'
    authoritative: false
  - name: X-Rate-Limit-Reset
    meaning: Static number of UTC epoch seconds remaining until the 60-second window resets. Legacy, static value.
    example: '20'
    authoritative: false
  - name: Retry-After
    meaning: Returned with 503 Service Unavailable, dictating when to attempt the call again.
    source: https://docs.fullcontact.com/docs/response-codes-errors

exhaustion:
  status: 429
  name: Too Many Requests
  condition: >-
    Returned only when the delay required to stay within the limit would exceed
    1000ms.
  retry_guidance: >-
    The 429 response is itself delayed by the maximum possible rate-limiting
    latency, so the request can be retried immediately and safely.

related_status_codes:
  - status: 403
    note: >-
      "Your API key is invalid, missing, or has exceeded its quota. Plans that
      have overages enabled will not receive a 403 when they exceed their
      allotted matches — they will only receive a 403 for exceeding rate limit
      quotas."
    source: https://docs.fullcontact.com/docs/response-codes-errors

mcp:
  note: >-
    The hosted MCP server enforces the same account limits. A successful
    fullcontact_enrich_person invocation counts as one person.enrich call. On
    exhaustion it returns a normalized envelope with code rate_limit_exceeded
    and a retry_after_seconds field.
  source: https://docs.fullcontact.com/docs/authentication

x-evidence:
  - fetched: '2026-08-14'
    url: https://docs.fullcontact.com/docs/rate-limiting.md
    http_status: 200