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: Dify
providerId: dify
created: '2026-05-04'
modified: '2026-09-06'
generated: '2026-09-06'
method: searched
source: >-
  https://dify.ai/pricing,
  https://docs.dify.ai/en/cloud/use-dify/knowledge/knowledge-request-rate-limit,
  https://docs.dify.ai/en/api-reference/guides/errors
description: >-
  Dify's published limits. Two independent mechanisms exist and they behave differently: a
  workspace-wide Knowledge Request rate limit measured per minute, and a plan-level API request
  quota measured per month on the free tier only. Neither is signalled with RateLimit-* response
  headers — Dify documents no rate-limit headers at all, so a client can only discover exhaustion
  from the status code and error code. This file replaces a 2026-05-04 scaffold whose headers and
  numbers were invented.
headers:
  limit: null
  remaining: null
  reset: null
  retryAfter: null
  policy: null
  note: >-
    Checked, nothing to record. The published error guide documents only status codes and error
    codes on exhaustion; no X-RateLimit-*, RateLimit-* or Retry-After header is documented anywhere
    in the API reference or the OpenAPI, and the OpenAPI's 429 responses declare no headers.
responseCodes:
  throttled: 429
  quotaExceeded: 429
  planLimitOnKnowledgeWrites: 403
error_codes:
  - code: too_many_requests
    status: 429
    meaning: Concurrency ceiling — too many simultaneous requests for the app right now.
    action: Back off and retry with exponential backoff.
  - code: rate_limit_error
    status: 429
    meaning: A Dify Cloud plan quota is exhausted (for example workflow executions).
    action: Retrying will not clear it; it resets with the quota period or a plan change.
  - code: forbidden
    status: 403
    meaning: >-
      On Dify Cloud, knowledge write endpoints return plan limits as 403 with the same code used for
      access restrictions.
    action: >-
      Read the message, not the code — a 403 switch on code alone cannot tell a plan limit from an
      access restriction.
limit_count: 5
limits:
  - name: Knowledge Request Rate Limit — Sandbox
    tier: Sandbox
    scope: workspace
    metric: knowledge_requests
    limit: 10
    timeFrame: minute
    docs: https://docs.dify.ai/en/cloud/use-dify/knowledge/knowledge-request-rate-limit
    applies_to:
      - Creating, deleting and updating knowledge bases
      - Uploading, deleting, updating, disabling, enabling, archiving and restoring documents
      - Pausing and resuming document processing
      - Adding, deleting, updating and bulk-importing segments
      - Hit tests
      - Querying knowledge from apps and workflows (multi-path recall counts as one request)
  - name: Knowledge Request Rate Limit — Professional
    tier: Professional
    scope: workspace
    metric: knowledge_requests
    limit: 100
    timeFrame: minute
    docs: https://docs.dify.ai/en/cloud/use-dify/knowledge/knowledge-request-rate-limit
  - name: Knowledge Request Rate Limit — Team
    tier: Team
    scope: workspace
    metric: knowledge_requests
    limit: 1000
    timeFrame: minute
    docs: https://docs.dify.ai/en/cloud/use-dify/knowledge/knowledge-request-rate-limit
  - name: API Rate Limit — Sandbox
    tier: Sandbox
    scope: workspace
    metric: api_requests
    limit: 5000
    timeFrame: month
    docs: https://dify.ai/pricing
  - name: API Rate Limit — Professional and Team
    tier: Professional, Team
    scope: workspace
    metric: api_requests
    limit: null
    timeFrame: month
    docs: https://dify.ai/pricing
    note: >-
      Published as "No Dify API Rate Limit". A concurrency ceiling still exists and surfaces as
      too_many_requests; its numeric value is not published.
self_hosted:
  note: >-
    Community and Enterprise self-hosted deployments run on the operator's own infrastructure; Dify
    publishes no rate limits for them.