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: Dimensions
providerId: dimensions
created: '2026-05-04'
modified: '2026-09-06'
generated: '2026-09-06'
method: searched
source: >-
  https://docs.dimensions.ai/dsl/usagepolicy.html and https://docs.dimensions.ai/dsl/api.html
  ("Technical Limits" / "Important" callouts), corroborated by
  https://github.com/digital-science/dimensions-analytics-mcp/blob/main/docs/REFERENCE.md
note: >-
  REPLACES a 2026-05-04 bulk-sweep scaffold that asserted a 10 req/min free tier,
  a 1,000 req/month quota and X-RateLimit-* response headers. None of that is
  published by Dimensions. The published limit is a single flat 30 requests per
  IP per minute, and the provider's own MCP client documentation states plainly
  that the API returns NEITHER Retry-After NOR X-RateLimit-* headers.
tags:
  - Analytics
  - Research
  - Rate Limiting
  - Quotas
description: >-
  Published rate limits and technical caps for the Dimensions Analytics API
  (DSL v2). The limits are edge-enforced, identical for all subscribers, and
  documented as not modifiable.
headers:
  limit: null
  remaining: null
  reset: null
  retryAfter: null
  policy: null
  note: >-
    No rate-limit signalling headers are returned. The Dimensions Analytics MCP
    reference states: "The Dimensions API enforces 30 requests per IP per minute
    at the edge. It does not return Retry-After or X-RateLimit-* headers." The
    official client therefore prevents 429s with a client-side sliding-window
    limiter rather than reading a runtime signal. This is a real agent-readiness
    gap: a consumer cannot discover its remaining budget from a response.
responseCodes:
  throttled: 429
  quotaExceeded: null
  serviceUnavailable: 500
limit_count: 1
limits:
  - tier: all
    name: Requests per IP per minute
    scope: ip
    metric: requests_per_minute
    limit: 30
    burst: null
    timeFrame: minute
    modifiable: false
    applies:
      - Dimensions Analytics API
      - Dimensions CRIS API
    source: https://docs.dimensions.ai/dsl/usagepolicy.html
    note: >-
      "max. 30 requests per IP address per minute". Stated on both the Reasonable
      Use page and the API Access page; the API Access page adds that the same
      limit covers the CRIS API. Applied per source IP, not per API key, so
      multiple keys behind one NAT share the budget.
technical_limits:
  - name: Items in an `in` filter clause
    limit: 400
    example: search publications where id in [...]
  - name: Boolean filter conditions
    limit: 100
    example: search publications where field1 = value and field2 = value
  - name: Boolean full-text clauses
    limit: 100
    example: 'search publications in authors for "\"Alan Turing\" OR \"Stephen Hawking\""' 
  - name: Records returned in a single query
    limit: 1000
  - name: Records returned for a single search via pagination (limit + skip)
    limit: 50000
    note: 50 pages of 1,000 records
  - name: Facet buckets returned
    limit: 1000
    note: No pagination possible on facets
technical_limits_source: https://docs.dimensions.ai/dsl/usagepolicy.html
token_limits:
  note: >-
    "Your subscription to use the API may include a defined number of tokens. Any
    tokens generated in excess of that number may be revoked without notice, or
    subject to additional charges." The number is set per subscription contract
    and is not published.
  source: https://docs.dimensions.ai/dsl/usagepolicy.html
reasonable_use:
  summary: >-
    The API is licensed for complex analytical tasks that cannot be done in the
    Dimensions platform. Building local copies of Dimensions data, using it as an
    alternative to the API, or using it for model training is prohibited.
    MCP-mediated access is allowed at Dimensions' discretion and is measured
    against the average of non-MCP users; unreasonable usage can result in
    suspension.
  source: https://docs.dimensions.ai/dsl/usagepolicy.html
client_side_mitigation:
  library: '@digital-science-dsl/dimensions-analytics-mcp'
  default_cap_per_minute: 30
  env: DIMENSIONS_RATE_LIMIT_PER_MINUTE
  retries: 3
  retry_env: DIMENSIONS_MAX_RETRIES
  backoff: exponential with jitter, base DIMENSIONS_RETRY_DELAY_MS (default 1000ms)
  surfaced_fields:
    - retryAfter
    - clientRateLimit.remaining
    - clientRateLimit.retryAfterMs
  source: https://github.com/digital-science/dimensions-analytics-mcp/blob/main/docs/REFERENCE.md