generated: '2026-08-13'
method: searched
source: https://reprise.zendesk.com/hc/en-us/articles/18940321925659-HTML-Environment-Data-API
limit_count: 2
name: Reprise Rate Limits
description: >
  Reprise publishes no request-rate limit, no rate-limit response headers and no documented
  status code on exhaustion for any of its three API surfaces. What it does publish, on both
  analytics Data APIs, is a hard 100MB response-size ceiling per call plus explicit guidance to
  pull narrow time windows on a schedule rather than issuing one wide backfill. Those are the two
  real, provider-published limits recorded here. Everything else is undocumented, and this file
  says so rather than inventing enterprise-tier defaults.
specificationVersion: "0.1"

rateLimits:
  - name: HTML Environment Data API — response size ceiling
    api: reprise:html-environment-data-api
    scope: per-request
    limitType: response-bytes
    limit: 104857600
    limit_human: 100MB
    window: per request
    burst: null
    guidance: >
      Reprise recommends pulling regularly from a restricted time period (for example one day at
      a time). Because paging is by time window only — start_timestamp / end_timestamp, no cursor
      or offset — the 100MB ceiling is effectively the page size.
    documentationUrl: https://reprise.zendesk.com/hc/en-us/articles/18940321925659-HTML-Environment-Data-API

  - name: Clone Environment Data API — response size ceiling
    api: reprise:data-api
    scope: per-request
    limitType: response-bytes
    limit: 104857600
    limit_human: 100MB
    window: per request
    burst: null
    guidance: Same 100MB ceiling and same narrow-window pull guidance as the HTML Environment pipe.
    documentationUrl: https://reprise.zendesk.com/hc/en-us/articles/28931374746907-Clone-Environment-Data-API

response_headers:
  published: false
  checked:
    - X-RateLimit-Limit
    - X-RateLimit-Remaining
    - X-RateLimit-Reset
    - RateLimit
    - RateLimit-Policy
    - Retry-After
  note: >
    No rate-limit headers are documented for the Data APIs or the MCP server, and none could be
    observed anonymously — every Data API call requires a per-portal token and every MCP call
    returns 401 without an OAuth bearer token. An agent has no runtime throttling signal to read
    on either surface.

status_on_exhaustion:
  documented: false
  note: No 429 contract, no Retry-After semantics published.

undocumented_surfaces:
  - api: reprise:mcp
    note: >
      No published tool-call rate limit, concurrency cap or quota for the MCP server. Reprise
      notes only that large operations (for example a PII scrub replacing the same name hundreds
      of times) may be run by the assistant as a background task or flagged for the HTML editor —
      a capability boundary rather than a rate limit.
    source: https://reprise.zendesk.com/hc/en-us/articles/51490067736603-Reprise-MCP-Starter-Prompt-Library
  - api: reprise:html-environment-data-api
    note: Request-rate limits are not published; contract terms are handled through the CSM.

authentication:
  - type: apiKey
    in: query
    param: token
    description: >
      Both Data APIs authenticate with an opaque token on the query string. The HTML Environment
      token is created in the portal under Settings → API Management; the Clone Environment token
      is issued by the customer's CSM.
  - type: oauth2
    description: >
      The MCP server uses OAuth 2.0 authorization code with PKCE; see
      authentication/reprise-authentication.yml.

statusPage: https://status.reprise.com/
supportUrl: https://www.reprise.com/contact

x-evidence:
  fetched: '2026-08-13'