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: Postmark
providerId: postmark
created: '2026-05-04'
# 2026-08-13: upgraded generated -> searched, and CORRECTED. The 2026-05-04 sweep
# asserted a `Retry-After` response header and a per-second request limit. Neither
# is documented anywhere on postmarkapp.com — see unverified_claims_removed below.
method: searched
modified: '2026-08-13'
reconciled: true
tags:
  - Rate Limiting
  - Email
description: >-
  Postmark publishes NO numeric API rate limit and NO rate-limit response
  headers. It documents that a 429 exists and that you should slow down, and it
  documents hard payload/recipient caps that behave like limits. The only
  numeric throttle Postmark states anywhere is a self-policed SMTP connection
  recommendation.
sources:
  - https://postmarkapp.com/developer/api/overview
  - https://postmarkapp.com/support/article/do-we-have-rate-limits-for-smtp
  - https://postmarkapp.com/llms.txt
  - https://postmarkapp.com/pricing

headers:
  request: []
  response: []
  retryAfter: null
  rateLimitRemaining: null
  rateLimitReset: null
  note: >-
    Probed the developer overview, the webhooks overview, the official-libraries
    page, the pricing page, the sandbox pages and llms.txt on 2026-08-13 —
    zero occurrences of Retry-After, RateLimit-*, or X-RateLimit-*. An agent
    that hits a 429 gets a status code and nothing else: no budget, no reset
    time, no recommended backoff. This is the single biggest runtime-semantics
    gap in the profile.
    https://postmarkapp.com/developer/user-guide/rate-limits returns HTTP 200
    but serves Postmark's "Oops, page not found" body — a soft 404, not a rate
    limits page.

responseCodes:
  throttled: 429
  throttledTitle: Rate Limit Exceeded
  throttledMeaning: >-
    "We have detected that you are making requests at a rate that exceeds
    acceptable use of the API. You should reduce the rate at which you query the
    API." — https://postmarkapp.com/developer/api/overview
  payloadTooLarge: 413

limits:
  - name: Batch Email API messages per call
    scope: request
    metric: messages_per_request
    limit: 500
    timeFrame: request
    enforced: true
    errorCode: 410
    source: https://postmarkapp.com/developer/api/overview
  - name: Recipients per message
    scope: request
    metric: recipients_per_message
    limit: 50
    timeFrame: request
    enforced: true
    note: To, Cc and Bcc combined.
    source: https://postmarkapp.com/llms.txt
  - name: Single message size
    scope: request
    metric: bytes_per_message
    limit: 10485760
    human: 10 MB including attachments
    timeFrame: request
    enforced: true
    httpStatus: 413
  - name: Batch/Bulk payload size
    scope: request
    metric: bytes_per_request
    limit: 52428800
    human: 50 MB
    timeFrame: request
    enforced: true
    httpStatus: 413
  - name: Body field size
    scope: request
    metric: bytes_per_field
    limit: 5242880
    human: 5 MB each for TextBody and HtmlBody
    timeFrame: request
    enforced: true
  - name: SMTP concurrent connections
    scope: client-ip
    metric: concurrent_connections
    limit: 10
    timeFrame: concurrent
    enforced: false
    note: >-
      Postmark's own words: "We recommend that customers limit the total number
      of connections to our SMTP endpoint to 10 concurrent connections, per IP"
      and "We do not have any systems in place to enforce these limits, we rely
      on customers to self-police and be good actors on our systems."
    source: https://postmarkapp.com/support/article/do-we-have-rate-limits-for-smtp
  - name: Monthly email volume
    scope: account
    metric: emails_per_month
    limit: plan-dependent
    timeFrame: month
    enforced: true
    note: >-
      100/month on Free with no overage permitted; paid tiers meter overage per
      1,000 emails. See plans/postmark-plans-pricing.yml.
  - name: Daily email limit
    scope: account
    metric: emails_per_day
    limit: unlimited
    timeFrame: day
    enforced: false
    note: Published as "Unlimited" on every tier on the pricing page.
  - name: Templates per server
    scope: server
    metric: templates
    limit: 100
    enforced: true
    errorCode: 1105
  - name: Message streams per server
    scope: server
    metric: message_streams
    limit: 10
    enforced: true
    errorCode: 1225
  - name: Tag length
    scope: request
    metric: characters
    limit: 1000
    enforced: true

api_request_rate:
  documented: false
  numeric_limit: null
  note: >-
    Postmark documents that an API rate limit exists and that exceeding it
    returns 429, but publishes no number for it — not per second, per minute or
    per hour, and not per token, server or account. Raising it is a support
    conversation.

unverified_claims_removed:
  - claim: 'headers.retryAfter: Retry-After'
    reason: >-
      Not documented on any Postmark page checked on 2026-08-13. Removed rather
      than carried forward, because a false header contract is worse for an
      agent than a recorded absence.
  - claim: 'Email API rate: requests_per_second, "soft (raise via support)"'
    reason: >-
      No requests-per-second limit is published anywhere. The "raise via
      support" part is plausible but unsourced, so the whole entry was dropped
      and replaced with api_request_rate above, which states the absence
      honestly.

limit_count: 11