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: Abstract API
providerId: abstract-api
created: '2026-05-04'
modified: '2026-08-29'
generated: '2026-08-29'
method: searched
source: >-
  https://docs.abstractapi.com/api/email-reputation (Response and error codes),
  https://docs.abstractapi.com/.well-known/agent-skills/abstractapi/skill.md,
  https://www.abstractapi.com/api/email-verification-validation-api#pricing,
  https://www.abstractapi.com/api/ip-geolocation-api#pricing
reconciled: true
tags:
  - Data Validation
  - Geolocation
  - Email Verification
  - Rate Limiting
description: >-
  Abstract enforces a per-API-key ceiling measured in requests per SECOND, plus a
  plan quota measured per month or per year. Both are scoped to one product's key —
  there is no account-level pooled limit. Exhausting the per-second ceiling returns
  429; exhausting the plan quota returns 422.
limit_count: 5
responseCodes:
  throttled: 429
  quotaExceeded: 422
  note: >-
    422 for quota and 429 for rate is unusual and worth coding for explicitly — many
    APIs use 429 for both. Abstract's own docs table and its published agent skill
    agree on the split.
response_headers:
  published: false
  x_ratelimit: false
  ratelimit_rfc9239: false
  retry_after: false
  finding: >-
    Abstract publishes NO rate-limit response headers of any kind. There is no
    X-RateLimit-Limit/Remaining/Reset, no RFC-draft RateLimit-*, and no Retry-After on
    429. A caller cannot read its remaining budget or a server-suggested backoff from
    any response; the only runtime signal is the status code after the limit is already
    hit. Confirmed by reading the docs error table, the provider's own agent skill, and
    all 13 provider-published OpenAPI definitions — none declares a rate-limit header.
limits:
  - name: Free tier rate ceiling — Email Validation / Email Reputation
    scope: api-key
    metric: requests_per_second
    limit: 3
    timeFrame: second
    source: https://www.abstractapi.com/api/email-verification-validation-api
  - name: Free tier rate ceiling — IP Geolocation
    scope: api-key
    metric: requests_per_second
    limit: 1
    timeFrame: second
    source: https://www.abstractapi.com/api/ip-geolocation-api
    note: >-
      The provider's own agent skill states "free plans: 1 req/sec" as a global rule.
      It is not global — it is the IP Geolocation figure. Email products allow 3 req/s
      on free.
  - name: Paid tier rate ceiling (Starter / Standard / Professional)
    scope: api-key
    metric: requests_per_second
    limit: 3
    timeFrame: second
    note: The same 3 req/s applies on every paid tier below Enterprise, across products.
  - name: Enterprise rate ceiling
    scope: api-key
    metric: requests_per_second
    limit: null
    timeFrame: second
    note: '"Flexible requests / second" — negotiated, not published.'
  - name: Plan quota
    scope: api-key
    metric: request
    limit: null
    timeFrame: month
    note: >-
      Per-product and per-tier; see plans/abstract-api-plans-pricing.yml. Free is 100
      (email) or 1,000 (IP geolocation) per month. Metered per REQUEST, not per
      successful response — a failed lookup still consumes quota.
scope_model:
  pooled: false
  note: >-
    Limits are per product key. Fifteen products means fifteen independent budgets;
    saturating one does not throttle another.
gaps:
  - No rate-limit response headers.
  - No Retry-After on 429.
  - Free-tier rate ceiling differs per product and is only discoverable on each product's pricing table.
  - No documented burst allowance.