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: X
providerId: x
created: '2026-05-04'
modified: '2026-08-28'
generated: '2026-08-28'
method: searched
source: https://docs.x.com/x-api/fundamentals/rate-limits.md
docs: https://docs.x.com/x-api/fundamentals/rate-limits
tags:
  - Rate Limiting
  - Quotas
  - Throttling
description: >-
  Published X API v2 rate limits, transcribed from X's own per-endpoint tables on
  2026-08-28. This file REPLACES a 2026-05-04 scaffold that carried invented tier numbers
  (10 rpm free / 100 rpm pro / 1000 rpm enterprise); X does not rate-limit by tier at all —
  limits are per-endpoint, per-app and per-user, on 15-minute or 24-hour windows.
model:
  dimension: per-endpoint
  scopes:
    - name: per-app
      description: Applies when calling with an app-only Bearer token.
    - name: per-user
      description: Applies when calling with an OAuth 1.0a or OAuth 2.0 user token.
  windows: ['15 minutes', '24 hours', '1 second']
  tiered: false
  note: >-
    Rate limits are independent of spend. Pay-per-usage adds a separate volume ceiling —
    3,000,000 Post reads per monthly billing cycle — which is a usage cap, not a rate limit.
  console: https://console.x.com
headers:
  limit: x-rate-limit-limit
  remaining: x-rate-limit-remaining
  reset: x-rate-limit-reset
  retryAfter: null
  policy: null
  reset_format: unix timestamp in seconds
  note: >-
    X does NOT send Retry-After and does NOT implement the IETF RateLimit-* draft headers.
    A client must read the absolute x-rate-limit-reset timestamp and compute its own wait.
responseCodes:
  throttled: 429
  quotaExceeded: 429
  serviceUnavailable: 503
  error_body:
    v2: '{"type":"https://api.x.com/2/problems/rate-limit-exceeded", ...}'
    legacy: '{"errors":[{"code":88,"message":"Rate limit exceeded"}]}'
limits:
  - name: Posts lookup by ids
    endpoint: GET /2/tweets
    perApp: 3500
    perUser: 5000
    timeFrame: 15 minutes
  - name: Post lookup by id
    endpoint: GET /2/tweets/:id
    perApp: 450
    perUser: 900
    timeFrame: 15 minutes
  - name: Recent search
    endpoint: GET /2/tweets/search/recent
    perApp: 450
    perUser: 300
    timeFrame: 15 minutes
    notes: 10 default / 100 max results; 512-character query length
  - name: Full-archive search
    endpoint: GET /2/tweets/search/all
    perApp: 300
    perUser: null
    timeFrame: 15 minutes
    burst: 1 request per second
    notes: 10 default / 500 max results; 1024-character query length; 1/sec per app and per user
  - name: Recent post counts
    endpoint: GET /2/tweets/counts/recent
    perApp: 300
    perUser: null
    timeFrame: 15 minutes
    notes: 512-character query length
  - name: Full-archive post counts
    endpoint: GET /2/tweets/counts/all
    perApp: 300
    perUser: null
    timeFrame: 15 minutes
    notes: 1024-character query length
  - name: Filtered stream connect
    endpoint: GET /2/tweets/search/stream
    perApp: 50
    perUser: null
    timeFrame: 15 minutes
    notes: 1 connection; 1000 rules; 1024-character rule length; 250 posts/sec delivery
  - name: Filtered stream rules lookup
    endpoint: GET /2/tweets/search/stream/rules
    perApp: 450
    perUser: null
    timeFrame: 15 minutes
  - name: Filtered stream rules update
    endpoint: POST /2/tweets/search/stream/rules
    perApp: 100
    perUser: null
    timeFrame: 15 minutes
  - name: Create Post
    endpoint: POST /2/tweets
    perApp: 10000
    perAppTimeFrame: 24 hours
    perUser: 100
    timeFrame: 15 minutes
  - name: Delete Post
    endpoint: DELETE /2/tweets/:id
    perApp: null
    perUser: 50
    timeFrame: 15 minutes
  - name: User timeline
    endpoint: GET /2/users/:id/tweets
    perApp: 10000
    perUser: 900
    timeFrame: 15 minutes
  - name: User mentions
    endpoint: GET /2/users/:id/mentions
    perApp: 450
    perUser: 300
    timeFrame: 15 minutes
  - name: Reverse chronological timeline
    endpoint: GET /2/users/:id/timelines/reverse_chronological
    perApp: null
    perUser: 180
    timeFrame: 15 minutes
  - name: Liking users lookup
    endpoint: GET /2/tweets/:id/liking_users
    perApp: 75
    perUser: 75
    timeFrame: 15 minutes
  - name: Liked posts lookup
    endpoint: GET /2/users/:id/liked_tweets
    perApp: 75
    perUser: 75
    timeFrame: 15 minutes
  - name: Like a Post
    endpoint: POST /2/users/:id/likes
    perApp: null
    perUser: 50
    timeFrame: 15 minutes
    dailyCap: 1000
    dailyTimeFrame: 24 hours
limit_count: 17
coverage_note: >-
  X publishes per-endpoint limits for roughly 25 Posts endpoints plus Users, Lists, Spaces,
  DMs, Media, Compliance and stream endpoints. The 17 rows above are the published Posts and
  timeline block, transcribed verbatim; the remainder are on the same page and in the
  Developer Console. Rows were not extrapolated to endpoints X did not publish a number for.
policies:
  - name: Backoff Strategy
    description: >-
      X's own guidance: on 429 read x-rate-limit-reset, implement exponential backoff, cache
      responses, and spread requests across the window.
    source: https://docs.x.com/x-api/fundamentals/rate-limits
  - name: Prefer streaming over polling
    description: >-
      X recommends the filtered stream rather than polling search endpoints for real-time data,
      explicitly as a rate-limit strategy.
  - name: Usage cap
    description: >-
      Pay-per-usage accounts are capped at 3,000,000 Post reads per monthly billing cycle.
      Exceeding it returns 429 with problem type usage-capped. Higher volume requires an
      Enterprise agreement.
    source: https://docs.x.com/x-api/getting-started/pricing
  - name: Separate limits per product
    description: >-
      X API v2, the Enterprise APIs and the Ads API each have their own rate-limit regimes;
      Enterprise limits are set by contract.
maintainers:
  - FN: Kin Lane
    email: kin@apievangelist.com