generated: '2026-09-12'
method: searched
source: https://devportal.gracenote.com/audio/gmd-api-v2/error-codes
provider: Gracenote
providerId: gracenote
description: >-
  Gracenote enforces rate limits and publishes the exhaustion behaviour, but does NOT publish the
  numeric limits — they are per-contract, set by the plan a customer licenses. No rate-limit response
  headers are declared in any of the ten public OpenAPI definitions, and none are documented.
limit_count: 0
limits: []
exhaustion:
  status: 429
  body_error_codes:
  - code: rate_limit_exceeded
    api: GMD API
    description: Number of queries per minute exceeds limit.
  - code: quota_limit_exceeded
    api: Nexus API
    description: Too many requests. Client exceeded their allocated rate limit.
  - code: plan_limit_exceeded
    status: 403
    api: GMD API
    description: Number of queries exceeds plan limit. Returned as 403, not 429 — a separate, plan-level ceiling.
  guidance: >-
    Gracenote's documented remediation is client-side: "Implement rate limiting and retry logic with
    exponential backoff."
headers:
  published: false
  retry_after: false
  ratelimit_draft: false
  x_ratelimit: false
  note: >-
    Probed the ten published OpenAPI definitions for X-RateLimit-*, RateLimit-* and Retry-After
    response headers. Zero occurrences. An agent gets no runtime budget signal from Gracenote — it
    learns it is throttled only by receiving a 429.
payload_limits:
- api: GMD API
  limit: 3MB request size
  error: request_too_large (413)
- api: GMD API
  limit: 100 IDs per request
  source: https://devportal.gracenote.com/audio/gmd-api-v2/error-codes
- api: GVD API
  limit: Gracenote recommends a maximum limit of 1000 across all endpoints
  source: openapi/gracenote-gvd-api-openapi.json info.description
- api: On API
  limit: Gracenote recommends a maximum limit of 1000 across all endpoints
  source: openapi/gracenote-on-api-openapi.json info.description
- api: Video MCP Server
  limit: >-
    Per-tool input ceilings — max 5 entities per gnv_resolve_entities call, max 1 TMSID per
    gnv_get_availability / gnv_get_tmsid / gnv_get_images / gnv_get_cast / gnv_get_awards /
    gnv_get_ratings call, max 3 actor names, max 400 characters per web-search query, 1-5 results.
  source: https://devportal.gracenote.com/video/video-mcp-server/tool-reference