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: Decart
providerId: decart
generated: '2026-08-01'
method: derived
created: '2026-08-01'
modified: '2026-08-01'
tags:
- Rate Limiting
- Real Time
- Artificial Intelligence
description: 'Decart does not publish per-second request rate limits. The throttling
  contract it does expose is a concurrency quota on realtime sessions, readable at
  runtime through GET /v1/realtime/quota, which returns the account limit, the currently
  active session count, and the remaining headroom. Client tokens can further constrain
  a single session with constraints.realtime.maxSessionDuration, and uploaded files
  carry a TTL. Values are account-specific and are not published as fixed numbers,
  so this catalog records the mechanism rather than fabricated limits.'
sources:
- openapi/decart-api-openapi-original.json
- https://docs.platform.decart.ai/getting-started/client-tokens
- https://docs.platform.decart.ai/models/realtime/streaming-best-practices
headers: {}
responseCodes:
  validation: 422
  unauthenticated: 401
  forbidden: 403
runtime_quota_endpoint:
  operationId: get_realtime_quota_v1_realtime_quota_get
  path: /v1/realtime/quota
  method: get
  returns:
    limit: Maximum concurrent realtime sessions for the account (nullable)
    active: Currently active realtime sessions
    remaining: Remaining concurrent realtime sessions
limits:
- name: Concurrent realtime sessions
  scope: account
  metric: concurrent_sessions
  limit: -1
  timeFrame: concurrent
  note: Account-specific; not published as a fixed number. Read live from GET /v1/realtime/quota.
- name: Client token allowedModels entries
  scope: token
  metric: models
  limit: 20
  timeFrame: request
  note: A client token may name at most 20 allowed models.
- name: Client token allowedOrigins entries
  scope: token
  metric: origins
  limit: 20
  timeFrame: request
  note: A client token may name at most 20 allowed web origins.
- name: Client token default lifetime
  scope: token
  metric: seconds
  limit: 60
  timeFrame: token
  note: Default expiresIn when the caller does not set one. Configurable per token.
- name: Uploaded file TTL
  scope: file
  metric: seconds
  limit: 2592000
  timeFrame: file
  note: TTL range is 60 to 2,592,000 seconds, or "persistent" for no expiry. Platform
    default when omitted is 24 hours.
notes:
- No 429 response is declared on any operation in either published OpenAPI document,
  and no Retry-After or X-RateLimit-* header is documented. Recorded as absent rather
  than assumed.
- The API Terms of Service prohibit circumventing rate limits and quotas, confirming
  they exist server-side even though the numbers are not published.