generated: '2026-08-13'
method: searched
source: https://github.com/vendasta/api-gateway-docs
notes: >-
  Vendasta publishes NO platform-wide API rate limit. The only numeric limits it documents are
  product-specific, and they are carried inside the OpenAPI operation descriptions of the AI
  Knowledge (embeddings) API rather than on a dedicated limits page. No rate-limit response headers
  (X-RateLimit-*, RateLimit-*, Retry-After) are declared anywhere in any of the 31 Vendasta OpenAPI
  documents in this repo, and no operation declares a 429 response — so an agent cannot read a
  remaining-quota signal at runtime and can only detect exhaustion from the bare HTTP 429 status.
  The webhook delivery timeouts and retry backoff below are the provider's outbound-direction
  limits, published in the Marketplace webhook reference; they are recorded here because they bound
  how fast an integration must respond, but they are not request rate limits.
limit_count: 2
response_headers: []
headers_documented: false
status_on_exhaustion: 429
limits:
- id: ai-knowledge-write-per-account-group
  api: AI Knowledge (embeddings)
  scope: per-account-group
  applies_to:
  - Knowledge_UpsertKnowledge
  - Knowledge_CreateKnowledgeFileUpload
  - Knowledge_RemoveKnowledge
  limit: 10
  unit: requests
  window: 1 minute
  refill: continuous
  burst: null
  status_on_exhaustion: 429
  source: openapi/vendasta-ai-knowledge-openapi.json
  quote: >-
    All write requests (Upsert Knowledge, Create Knowledge File Upload, and Remove Knowledge
    together) share one rate limit. The write budget of each account group refills continuously,
    at 10 requests per minute.
- id: ai-knowledge-write-partner-ceiling
  api: AI Knowledge (embeddings)
  scope: per-partner
  applies_to:
  - Knowledge_UpsertKnowledge
  - Knowledge_CreateKnowledgeFileUpload
  - Knowledge_RemoveKnowledge
  limit: 100
  unit: requests
  window: 1 minute
  burst: null
  status_on_exhaustion: 429
  source: openapi/vendasta-ai-knowledge-openapi.json
  quote: >-
    A partner-wide ceiling of 100 requests per minute applies to all account groups together. A
    request over either limit fails with HTTP 429. Writes to the partner's own knowledge base count
    only against the partner-wide ceiling.
webhook_delivery:
  note: >-
    Outbound direction — how long the consumer has to respond before Vendasta retries. Published in
    the Marketplace webhook reference.
  source: https://github.com/vendasta/marketplace-documentation/blob/master/docs/Other/marketplace_webhooks.md
  rules:
  - {webhook: Purchase webhooks, timeout: 30s, retry: backoff doubles from 30s to a maximum spacing of one hour}
  - {webhook: All other Marketplace webhooks, timeout: 10s, retry: backoff doubles from 60s to a maximum spacing of one hour}
  - {webhook: Automation "Send a webhook" action, timeout: 60s, retry: resent when response exceeds 60s, HTTP >= 500, or the connection breaks}
gaps:
- No published rate limit for the Platform REST APIs, CRM, SCIM, Reputation, Social, Local SEO, Advertising or Customer Voice APIs.
- No rate-limit response headers declared in any spec; exhaustion is only observable as HTTP 429.
- No documented Retry-After on 429.
checked: '2026-08-13'