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: Eden AI
providerId: eden-ai
created: '2026-05-04'
modified: '2026-09-06'
generated: '2026-09-06'
method: searched
source: https://www.edenai.co/docs/v3/overview/rate-limits
docs:
  - https://www.edenai.co/docs/v3/overview/rate-limits
  - https://www.edenai.co/docs/v3/overview/plans-prices
tags:
  - Artificial Intelligence
  - AI Gateway
  - Rate Limiting
  - Quotas
description: >-
  Published Eden AI rate limits, read from the provider's own documentation on 2026-09-06. This file
  REPLACES a 2026-05-04 scaffold that carried invented headers and per-tier quotas; none of those
  values were ever published by Eden AI.
limits:
  - id: self-serve-default
    scope: per-account
    applies_to: 'All endpoints under https://api.edenai.run/v3/'
    window: 1s
    limit: 10
    unit: requests
    burst: null
    plan: Self-serve AI API
    adjustable: true
    adjustment_process: >-
      Raised on request via the live chat with a description of the use case and expected throughput.
    source: https://www.edenai.co/docs/v3/overview/rate-limits
    note: >-
      Shared across all of an account's API keys. In an organization the owner's limit is the ceiling
      for every member, and the owner can cap individual members below it.
  - id: advanced-platform
    scope: per-account
    window: null
    limit: null
    unit: requests
    plan: Advanced AI Platform
    adjustable: true
    source: https://www.edenai.co/docs/v3/overview/plans-prices
    note: >-
      "Custom (higher)" — tailored to the workload, arranged with sales. No number is published, so
      none is recorded.
documentation_conflict:
  observed: '2026-09-06'
  detail: >-
    Eden AI's two published statements of the Self-serve limit disagree. /docs/v3/overview/rate-limits
    says "Every account starts at 10 requests per second" and the plans table says "10 reqs/sec
    (+ on request)", while the Self-serve bullet list on the same plans page says "7 requests/second
    by default, upgradable to 15 req/sec on request". Recorded verbatim rather than reconciled — an
    integrator cannot tell which is authoritative either.
  sources:
    - https://www.edenai.co/docs/v3/overview/rate-limits
    - https://www.edenai.co/docs/v3/overview/plans-prices
headers:
  documented: false
  note: >-
    Eden AI does not document any rate-limit response headers (no X-RateLimit-*, no RateLimit-*, no
    Retry-After) and none appear in any of the twelve first-party OpenAPI definitions. An agent
    cannot read remaining budget or a reset time from a response; it can only back off blindly. The
    previous scaffold in this file asserted a full X-RateLimit-* header set that Eden AI has never
    published — that claim has been removed.
responseCodes:
  throttled: null
  quotaExceeded: 402
  note: >-
    429 is named as a retriable HTTP-level failure in Eden AI's own published Agent Skill
    (skills/eden-ai-edenai.md, "Retry with backoff on 429/5xx"), but no rate-limit status code is
    stated in the rate-limit documentation itself, so it is not recorded as published. A depleted
    credit balance is documented as 402 Payment Required
    (https://www.edenai.co/docs/v3/general/buying-credits).
spend_controls:
  - {control: per-key balance, fields: [balance, balance_reset_period, balance_reset_amount], docs: 'https://www.edenai.co/docs/v3/general/custom-api-keys'}
  - {control: guardrails, description: 'Reusable usage policies controlling which models a team can call, how often, and how much a member can spend', docs: 'https://www.edenai.co/docs/v3/organization/guardrails'}
  - {control: postpaid monthly spend limit, description: 'Optional cap; further requests blocked until the next billing period', docs: 'https://www.edenai.co/docs/v3/general/buying-credits'}
limit_count: 2