specification: API Commons Rate Limits
specificationVersion: '0.1'
schema: https://raw.githubusercontent.com/api-evangelist/interface-research/main/schema/api-commons.yml#/$defs/RateLimits
provider: Brand API (Brandfetch)
providerId: brand-api
created: '2026-05-04'
# Provenance stamped 2026-08-11: this artifact was written by the API Evangelist
# bulk sweep dated 2026-05-04, not harvested from the provider. See roadmap#35.
# UPGRADED 2026-08-14: every number below was re-read against the provider's live
# rate-limit documentation and confirmed; the response headers, the Brand Search API
# per-IP limit and the 429 message were added from the same pass. method: generated ->
# searched.
method: searched
source: https://docs.brandfetch.com/logo-api/overview
modified: '2026-08-14'
reconciled: true
tags:
  - Brands
  - Logos
  - Brand Assets
  - Rate Limiting
description: Brandfetch enforces a soft 500,000 requests/month fair-use ceiling on the Logo API plus
  hard 5-minute throughput windows scoped per IP and per customer. Overages on the throughput windows
  return HTTP 429. The Brand API has a separate hard cap of 2,500 calls/month tied to its subscription.
sources:
  - https://docs.brandfetch.com/logo-api/overview
  - https://docs.brandfetch.com/.well-known/agent-skills/brandfetch/skill.md
  - openapi/brand-api-brandfetch-openapi.yml
responseCodes:
  throttled: 429
  throttledMessage: API key quota exceeded
responseHeaders:
  - name: x-api-key-quota
    meaning: Quota allotted to the API key.
    source: https://docs.brandfetch.com/.well-known/agent-skills/brandfetch/skill.md
  - name: x-api-key-approximate-usage
    meaning: Approximate usage consumed against that quota.
    source: https://docs.brandfetch.com/.well-known/agent-skills/brandfetch/skill.md
  - name: Retry-After
    meaning: Not documented and not observed. Do not depend on it.
    present: false
headerNote: >-
  Brandfetch publishes no standard RateLimit-* (RFC 9239 draft) or X-RateLimit-* headers.
  The two x-api-key-* headers above are the only runtime quota signal, and they are named
  in Brandfetch's own published Agent Skill verification checklist rather than in the API
  reference. An agent should read them on every response; there is no header that reports
  the 5-minute throughput window.
limits:
  - name: Logo API monthly fair-use
    scope: account
    metric: requests_per_month
    limit: 500000
    timeFrame: month
    notes: Soft limit. Exceeding does not trigger immediate blocking.
  - name: Logo API per-IP throughput
    scope: ip
    metric: requests_per_5_minutes
    limit: 1000
    timeFrame: minute
    notes: Hard 1,000 requests every 5 minutes per IP. 429 on overage.
  - name: Logo API per-customer throughput
    scope: account
    metric: requests_per_5_minutes
    limit: 2400
    timeFrame: minute
    notes: Hard 2,400 requests every 5 minutes per customer. 429 on overage.
  - name: Brand Search API per-IP throughput
    scope: ip
    metric: requests_per_5_minutes
    limit: 200
    timeFrame: minute
    notes: >-
      200 requests every 5 minutes per IP on the Brand Search API. Brandfetch's published
      Agent Skill tells implementers to debounce autocomplete inputs because of it.
      Confirmed 2026-08-14 from
      https://docs.brandfetch.com/.well-known/agent-skills/brandfetch/skill.md
  - name: Brand API monthly cap
    scope: account
    metric: requests_per_month
    limit: 2500
    timeFrame: month
    notes: >-
      Hard cap on the $99/month Brand API subscription, per the 2026-05 pricing sweep.
      NOT re-verified on 2026-08-14 — brandfetch.com/developers/pricing returns HTTP 403
      to every non-browser client. The free developer plan's 100 requests/month IS
      confirmed (https://docs.brandfetch.com/mcp/overview).
  - name: Free developer plan monthly quota
    scope: account
    metric: requests_per_month
    limit: 100
    timeFrame: month
    notes: >-
      Confirmed 2026-08-14 from https://docs.brandfetch.com/mcp/overview — "sign up for
      free; the free plan includes 100 requests per month." MCP calls draw on the same
      quota as REST calls.
policies:
  - name: Backoff Strategy
    description: Apply exponential backoff with jitter on 429 responses; respect any Retry-After header
      when present.
  - name: Soft vs Hard Limits
    description: The 500K monthly Logo API allowance is soft (fair-use); per-IP and per-customer 5-minute
      throughput windows are hard limits.
  - name: Caching
    description: Cache logo and brand responses aggressively client-side to stay inside the 5-minute
      throughput windows; Enterprise plans negotiate custom cache TTLs.
  - name: Capacity Increase
    description: Higher monthly volumes and SLAs require an Enterprise plan negotiated with Brandfetch.
maintainers:
  - FN: Kin Lane
    email: info@apievangelist.com