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: MATLAB
providerId: matlab
created: '2026-05-04'
modified: '2026-09-14'
generated: '2026-09-14'
method: searched
source: >-
  https://thingspeak.mathworks.com/pages/license_faq,
  https://thingspeak.mathworks.com/prices,
  https://www.mathworks.com/help/thingspeak/error-codes.html,
  https://www.mathworks.com/help/thingspeak/get-a-channel-feed.html,
  https://www.mathworks.com/help/thingspeak/mqtt-basics.html
note: >-
  Replaces the 2026-05-04 generated scaffold, which invented X-RateLimit-* headers, a 10 req/min free
  tier and a 1,000 req/month quota and applied them to all thirteen apis[] entries. None of that was
  published by MathWorks. The real published limits are ThingSpeak's, they are expressed as a minimum
  interval between channel updates plus an annual message allowance, and MathWorks documents NO
  rate-limit response headers at all. Most of the apis[] entries are in-process language bindings with
  no HTTP surface and therefore no rate limit.
headers:
  limit: null
  remaining: null
  reset: null
  retryAfter: null
  policy: null
headers_note: >-
  Checked, nothing to record. No X-RateLimit-*, RateLimit-* or Retry-After header is documented on any
  MathWorks HTTP surface. A client discovers throttling only from the status code, which means an agent
  cannot pace itself — it can only react after being refused.
responseCodes:
  throttled: 429
  quotaExceeded: 402
  note: >-
    ThingSpeak splits these deliberately: 429 error_too_many_requests means "you called too fast, back
    off", 402 error_payment_required means "your licence's annual message allowance is spent". Same
    family of failure, completely different remedy.
limits:
  - id: thingspeak-free-update-interval
    api: ThingSpeak REST API
    tier: free
    name: Channel update interval (free licence)
    scope: per-channel
    metric: seconds_between_updates
    limit: 15
    unit: seconds
    source: https://thingspeak.mathworks.com/pages/license_faq
  - id: thingspeak-paid-update-interval
    api: ThingSpeak REST API
    tier: paid
    name: Channel update interval (paid licence)
    scope: per-channel
    metric: seconds_between_updates
    limit: 1
    unit: seconds
    source: https://thingspeak.mathworks.com/prices
  - id: thingspeak-free-messages
    api: ThingSpeak REST API
    tier: free
    name: Annual message allowance (free licence)
    scope: per-account
    metric: messages
    limit: 3000000
    timeFrame: year
    source: https://thingspeak.mathworks.com/pages/license_faq
  - id: thingspeak-paid-messages
    api: ThingSpeak REST API
    tier: paid
    name: Annual message allowance (per purchased unit)
    scope: per-account
    metric: messages
    limit: 33000000
    timeFrame: year
    source: https://thingspeak.mathworks.com/prices
  - id: thingspeak-free-storage
    api: ThingSpeak REST API
    tier: free
    name: Stored message ceiling (free licence)
    scope: per-account
    metric: messages_stored
    limit: 10000000
    source: https://thingspeak.mathworks.com/pages/license_faq
  - id: thingspeak-free-channels
    api: ThingSpeak REST API
    tier: free
    name: Channels (free licence)
    scope: per-account
    metric: channels
    limit: 4
    source: https://thingspeak.mathworks.com/pages/license_faq
  - id: thingspeak-paid-channels
    api: ThingSpeak REST API
    tier: paid
    name: Channels (per purchased standard unit)
    scope: per-account
    metric: channels
    limit: 250
    source: https://thingspeak.mathworks.com/prices
  - id: thingspeak-read-results
    api: ThingSpeak REST API
    tier: all
    name: Maximum entries per read
    scope: per-request
    metric: results
    limit: 8000
    source: https://www.mathworks.com/help/thingspeak/get-a-channel-feed.html
    note: Hard ceiling on the results parameter. There is no cursor, so larger reads must page by moving the start/end window.
  - id: thingspeak-mqtt-channels-per-device
    api: ThingSpeak MQTT API
    tier: all
    name: Authorized channels per MQTT device
    scope: per-device
    metric: channels
    limit: 10
    source: https://www.mathworks.com/help/thingspeak/mqtt-basics.html
  - id: thingspeak-mqtt-devices
    api: ThingSpeak MQTT API
    tier: all
    name: MQTT devices per user
    scope: per-account
    metric: devices
    limit: 100
    source: https://www.mathworks.com/help/thingspeak/mqtt-basics.html
  - id: thingspeak-mqtt-idle-timeout
    api: ThingSpeak MQTT API
    tier: all
    name: MQTT connection idle timeout
    scope: per-connection
    metric: minutes_idle
    limit: 60
    source: https://www.mathworks.com/help/thingspeak/mqtt-basics.html
billing_unit:
  name: message
  definition: >-
    "A write of up to 8 fields of data to a ThingSpeak channel." REST writes, MQTT writes, MATLAB
    Analysis outputs and React/TimeControl triggers all count. An image write costs 100 messages.
  source: https://thingspeak.mathworks.com/pages/license_faq
undocumented:
  - api: MATLAB Production Server RESTful API
    reason: >-
      Customer-hosted. Throughput is bounded by the worker pool the operator configures
      (see MATLABProductionServerSizer), not by a MathWorks-published limit.
  - api: MATLAB Engine API for Python / Java / C++ / C and Fortran / .NET
    reason: In-process language bindings. No HTTP surface, therefore no rate limit.
  - api: MATLAB MEX API
    reason: In-process. No HTTP surface.
  - api: MATLAB Data API for C++
    reason: In-process. No HTTP surface.
  - api: MATLAB Compiler SDK API
    reason: Build-time API. No HTTP surface.
  - api: MATLAB HTTP Interface
    reason: A client library for calling other people's APIs; it is not itself a rate-limited service.
  - api: MATLAB Web App Server
    reason: Customer-hosted. No MathWorks-published limit.
limit_count: 11