generated: '2026-08-13'
method: searched
source: https://docs.lightfield.app/using-the-api/rate-limits/
docs: https://docs.lightfield.app/using-the-api/rate-limits/
checked: '2026-08-13'
api: https://api.lightfield.app/v1
summary: >-
  Lightfield publishes a flat 25 requests-per-second ceiling per organization, applied separately to
  three method categories (write, read, search). Limits are described as early-launch values subject
  to change. Every response carries X-RateLimit-* headers and exhaustion returns 429 with
  Retry-After, so the runtime signal is available to an agent without reading the docs.
scope: per organization
limit_count: 3
limits:
- category: Write
  methods: [Create, Update]
  limit: 25
  unit: requests
  window: 1s
  scope: organization
- category: Read
  methods: [Retrieve, Definitions]
  limit: 25
  unit: requests
  window: 1s
  scope: organization
- category: Search
  methods: [List]
  limit: 25
  unit: requests
  window: 1s
  scope: organization
headers:
- name: X-RateLimit-Limit
  description: Maximum requests allowed per second (bucket capacity)
- name: X-RateLimit-Remaining
  description: Number of requests remaining in the current window
- name: X-RateLimit-Reset
  description: Unix timestamp (seconds) when the bucket will be fully replenished
standard_headers: false
standard_headers_note: >-
  Lightfield uses the legacy X-RateLimit-* convention, not the IETF RateLimit / RateLimit-Policy
  draft headers.
exhaustion:
  status: 429
  reason: Too Many Requests
  retry_after_header: Retry-After
  retry_after_units: seconds
  error_body:
    type: too_many_requests
    message: Rate limit exceeded. Please try again later.
  note: >-
    The 429 example published in the docs shows X-RateLimit-Limit: 5, which does not match the
    documented 25 rps ceiling — treat the header as the authoritative runtime value and the table as
    the launch-period default.
best_practices:
- Implement exponential backoff keyed off Retry-After.
- Monitor X-RateLimit-Remaining and throttle before exhaustion.
- Cache reads to avoid redundant GETs.
higher_limits:
  self_serve: false
  process: Contact the Lightfield team for higher throughput.
stability: >-
  "These limits are subject to change as the platform scales." Published as early-launch values
  while the API is in public beta.
quota_vs_rate:
  note: >-
    Rate limits are separate from commercial metering. Lightfield's plans meter consumption in
    annual credits (plans/lightfield-plans-pricing.yml); the docs do not state whether direct API
    calls draw down credits.
related:
  conventions: conventions/lightfield-conventions.yml
  errors: errors/lightfield-problem-types.yml
  plans: plans/lightfield-plans-pricing.yml