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: Segmind
providerId: segmind
generated: '2026-08-27'
method: searched
source: >-
  https://docs.segmind.com/docs/serverless-api/rate-limits,
  https://www.segmind.com/pricing,
  https://docs.segmind.com/docs/get-started/authentication,
  https://docs.segmind.com/docs/platform/account-and-billing-apis
created: '2026-05-04'
modified: '2026-08-27'
tags:
  - Rate Limiting
  - Quotas
  - Throttling
description: >-
  Published rate limits for the Segmind AI Gateway, harvested from the provider's own
  rate-limits documentation, pricing page and authentication guide on 2026-08-27. This
  replaces the 2026-05-04 scaffold, whose tier numbers were placeholders.
limit_count: 7
conflict:
  status: unresolved
  summary: >-
    Segmind publishes two different per-minute rate-limit tables and they do not agree.
    The docs rate-limits page gives Free 5 RPM / Pro 50 RPM / Enterprise custom; the
    pricing page gives Flexible 60 RPM / Pro 120 RPM / Business 500 RPM / Scale 1000 RPM
    pooled. The plan names also differ — the docs table has no Business or Scale row at
    all, and the pricing page has no "Free" plan (the entry tier is Flexible, shown as
    "Free Plan" in the console until you subscribe). Both figures are recorded below
    against the page that publishes them. An integrator cannot tell from Segmind's public
    surface which one the gateway enforces.
  sources:
    - https://docs.segmind.com/docs/serverless-api/rate-limits
    - https://www.segmind.com/pricing
headers:
  reset: x-rate-limit-reset-at-utc
  note: >-
    The only rate-limit header Segmind documents is `x-rate-limit-reset-at-utc`, returned
    on v1 synchronous responses and described as "when your current rate-limit window
    resets". No X-RateLimit-Limit / X-RateLimit-Remaining / RateLimit-Policy / Retry-After
    header is documented, and the v2 async `metrics` object carries cost and timing but no
    rate-limit fields. An agent therefore cannot read its remaining budget from a response
    — only when the window resets.
  limit: null
  remaining: null
  retryAfter: null
  policy: null
responseCodes:
  throttled: 429
  quotaExceeded: 406
  note: >-
    429 `rate_limit_exceeded` on throttling. Running out of credits is a separate
    condition and returns 406 with an "Insufficient credits" error naming both the balance
    and what the call needed — not 429.
limits:
  - tier: free
    plan: Free
    name: Docs rate-limits page — Free
    scope: api-key
    metric: requests_per_minute
    limit: 5
    timeFrame: minute
    source: https://docs.segmind.com/docs/serverless-api/rate-limits
  - tier: professional
    plan: Pro
    name: Docs rate-limits page — Pro
    scope: api-key
    metric: requests_per_minute
    limit: 50
    timeFrame: minute
    source: https://docs.segmind.com/docs/serverless-api/rate-limits
  - tier: free
    plan: Flexible
    name: Pricing page — Flexible (pay as you go)
    scope: account
    metric: requests_per_minute
    limit: 60
    timeFrame: minute
    source: https://www.segmind.com/pricing
  - tier: professional
    plan: Pro
    name: Pricing page — Pro
    scope: account
    metric: requests_per_minute
    limit: 120
    timeFrame: minute
    source: https://www.segmind.com/pricing
  - tier: business
    plan: Business
    name: Pricing page — Business
    scope: account
    metric: requests_per_minute
    limit: 500
    timeFrame: minute
    source: https://www.segmind.com/pricing
  - tier: scale
    plan: Scale
    name: Pricing page — Scale (pooled)
    scope: account
    metric: requests_per_minute
    limit: 1000
    timeFrame: minute
    pooled: true
    source: https://www.segmind.com/pricing
  - tier: enterprise
    plan: Enterprise
    name: Enterprise — custom
    scope: contract
    metric: requests_per_minute
    limit: null
    timeFrame: minute
    note: Published as "Custom" on both the docs table and the pricing page.
    source: https://www.segmind.com/pricing
auth_surface_limits:
  - name: Authentication endpoints
    scope: ip
    metric: requests_per_minute
    limit: 5
    timeFrame: minute
    source: https://docs.segmind.com/docs/get-started/authentication
  - name: Token refresh
    scope: user
    metric: requests_per_hour
    limit: 10
    timeFrame: hour
    source: https://docs.segmind.com/docs/get-started/authentication
policies:
  - name: Retry guidance
    description: >-
      The authentication docs tell clients to implement retry with exponential backoff and
      to handle token expiration gracefully. No Retry-After header is documented to back
      that up, so the backoff is entirely client-side.
    source: https://docs.segmind.com/docs/get-started/authentication
  - name: Dedicated endpoints bypass the shared limit
    description: >-
      Dedicated GPU endpoints are billed per GPU-hour or per GPU-second rather than per
      call, and the pricing page describes them as "Unlimited API requests with dedicated
      endpoints". They are a separate control plane on api.spotprod.segmind.com.
    source: https://www.segmind.com/pricing
  - name: Why limits exist
    description: >-
      Segmind states the purpose as security (abuse mitigation), fair access (no single
      consumer monopolising capacity), and performance optimisation under spikes.
    source: https://docs.segmind.com/docs/serverless-api/rate-limits
maintainers:
  - FN: Kin Lane
    email: kin@apievangelist.com