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: Tray.ai
providerId: tray-ai
created: '2026-05-04'
# Provenance stamped 2026-08-11: this artifact was written by the API Evangelist
# bulk sweep dated 2026-05-04, not harvested from the provider. See roadmap#35.
method: searched
modified: '2026-09-02'
tags:
  - Automation
  - Integration
  - iPaaS
  - Rate Limiting
  - Quotas
  - Throttling
sourceUrls:
  - https://tray.ai/documentation/developer/getting-started/troubleshooting/rate-limiting-tray
  - https://tray.ai/documentation/developer/getting-started/troubleshooting/rate-limiting-3rd-party
description: >-
  Machine-readable rate limit definitions for Tray.ai's Platform API and Embedded API. Values
  reflect the limits documented in Tray's developer portal. Verbatim from the Platform API docs:
  "all endpoints are rate limited at 30 requests per second or 1800 requests per minute." The
  Embedded API docs match: "All endpoints are rate limited to 30 requests per second." The Call
  Connector endpoint is additionally capped at 1,000 concurrent requests. Trigger API event
  delivery is not rate limited.
# CORRECTION 2026-09-02: the previous version of this file listed X-RateLimit-Limit,
# X-RateLimit-Remaining, X-RateLimit-Reset, Retry-After and RateLimit-Policy as though Tray
# returned them. Tray documents none of these anywhere, and a live unauthenticated request to
# https://api.tray.io/core/v1/connectors (probed 2026-09-02, HTTP 401) returned no rate-limit
# header of any kind. The claim was a template default, not an observation, and it is removed.
headersPublished: false
headers: {}
headersNote: >-
  Tray publishes the NUMBERS but not the SIGNAL. There are no RateLimit-* / X-RateLimit-* headers
  and no Retry-After, so a client cannot read its remaining budget at runtime — it must model the
  published 30 rps / 1800 rpm ceiling and back off on its own schedule. Verified against the
  documentation and against a live response on 2026-09-02.
responseCodes:
  throttled: 429
  quotaExceeded: 429
evidence:
  - url: https://tray.ai/documentation/developer/getting-started/troubleshooting/rate-limiting-tray
    status: 200
    finding: >-
      "all endpoints (Embedded APIs/Connectivity APIs) are rate limited at 30 requests per second
      or 1800 requests per minute"; "limited burst capacity ... (up to 50 requests per second)";
      Call Connector "limits the concurrent calls to 1000"; "the delivery of events to your
      endpoint is not rate limited ... Tray will retry the delivery of those events with
      exponential backoff."
  - url: https://api.tray.io/core/v1/connectors
    status: 401
    finding: 'No RateLimit-* / X-RateLimit-* / Retry-After response header present.'
limits:
  - name: Platform API Steady-State
    description: All Tray Platform API (Connectivity API) endpoints are rate limited at 30 requests
      per second, equivalent to 1,800 requests per minute.
    api: Tray.ai Platform API
    scope: account
    metric: requests_per_second
    limit: 30
    timeFrame: second
    applies:
      - https://api.tray.io/core/v1
      - https://api.eu1.tray.io/core/v1
      - https://api.ap1.tray.io/core/v1
  - name: Platform API Burst
    description: Short bursts up to 50 requests per second are tolerated before throttling engages.
    api: Tray.ai Platform API
    scope: account
    metric: requests_per_second
    limit: 50
    timeFrame: second
    burst: true
  - name: Platform API Per-Minute Equivalent
    description: 30 RPS expressed as a per-minute ceiling per the published documentation.
    api: Tray.ai Platform API
    scope: account
    metric: requests_per_minute
    limit: 1800
    timeFrame: minute
  - name: Call Connector Concurrency
    description: The Call Connector endpoint has a 1,000 concurrent-requests ceiling in addition to
      the per-second rate limit.
    api: Tray.ai Platform API
    scope: account
    metric: concurrent_requests
    limit: 1000
    timeFrame: instant
    applies:
      - POST /connectors/{connector-name}/versions/{connector-version}/call
  - name: Trigger Event Delivery
    description: Trigger API event delivery is not rate limited; events are delivered as quickly as
      possible.
    api: Tray.ai Platform API
    scope: account
    metric: events
    limit: -1
    timeFrame: instant
  - name: Embedded API Steady-State
    description: All Tray Embedded GraphQL API endpoints are rate limited to 30 requests per second.
    api: Tray.ai Embedded API
    scope: account
    metric: requests_per_second
    limit: 30
    timeFrame: second
    applies:
      - https://tray.io/graphql
      - https://eu1.tray.io/graphql
      - https://ap1.tray.io/graphql
  - name: Embedded API Burst
    description: Short bursts up to 50 requests per second are tolerated on the Embedded API before
      throttling engages.
    api: Tray.ai Embedded API
    scope: account
    metric: requests_per_second
    limit: 50
    timeFrame: second
    burst: true
policies:
  - name: Backoff Strategy
    description: Clients should implement exponential backoff with jitter and honor the Retry-After
      header when present in a 429 response.
  - name: Burst Handling
    description: Short bursts up to 50 RPS are tolerated before steady-state throttling engages on
      both the Platform and Embedded APIs.
  - name: Concurrency
    description: The Call Connector endpoint enforces a 1,000-concurrent-request ceiling; long-running
      connector calls should be parallelized within that ceiling.
  - name: Event Delivery
    description: Trigger event delivery is not rate limited and is best-effort real-time. Consumers
      must size webhook receivers accordingly.
  - name: Fair Use
    description: Even though Trigger event delivery is uncapped, sustained patterns that materially
      impact shared infrastructure may be subject to fair-use review by Tray.
maintainers:
  - FN: Kin Lane
    email: kin@apievangelist.com