specification: API Commons Rate Limits
specificationVersion: '0.1'
provider: LlamaParse
providerId: llamaparse
created: '2026-06-12'
modified: '2026-06-12'
url: https://developers.llamaindex.ai/llamaparse/general/rate_limits/

throttled:
  statusCode: 429
  description: >
    When a rate limit is exceeded the API returns HTTP 429 Too Many Requests. Clients should
    implement exponential backoff or request batching. For persistent rate limit issues contact
    support@runllama.ai.
  retryAfter:
    header: Retry-After
    description: Standard HTTP Retry-After header may be returned; documentation does not guarantee its presence on all 429 responses.

limits:
  - scope: per_project
    metric: file_upload_requests
    limit: 50
    timeFrame: 5s
    description: File upload endpoint — 50 requests per 5-second window per project

  - scope: per_organization
    metric: parse_upload_requests
    limit: 50
    timeFrame: 10s
    description: Parse upload endpoint — 50 requests per 10-second window per organization

  - scope: per_organization
    metric: classify_requests
    limit: 40
    timeFrame: 1s
    description: Classify endpoint — 40 requests per second per organization

  - scope: per_organization_free_tier
    metric: requests_per_minute
    limit: 20
    timeFrame: 1m
    description: Free-tier organizations are subject to a global 20 requests-per-minute cap

tiers:
  - plan: Free
    requestsPerMinute: 20
    concurrentJobs: 5
    notes: Stricter rate limits apply to free-tier organizations across all endpoints

  - plan: Starter
    concurrentJobs: 5
    notes: Standard rate limits; no explicit RPM cap documented beyond per-endpoint QPS limits

  - plan: Pro
    concurrentJobs: 20
    notes: Standard rate limits with higher concurrency

  - plan: Enterprise
    concurrentJobs: 100
    notes: Custom rate limits available; contact account manager

contactForIncreases: support@runllama.ai