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: LlamaIndex
providerId: llamaindex
created: '2026-05-04'
# Provenance stamped 2026-08-11: this artifact was written by the API Evangelist
# bulk sweep dated 2026-05-04, not harvested from the provider. See roadmap#35.
method: generated
modified: '2026-05-05'
reconciled: false
tags:
  - Rate Limiting
  - LLM
  - RAG
description: LlamaIndex / LlamaCloud rate limits are not publicly enumerated as per-second numbers on
  the pricing page; tiered limits scale with plan and Enterprise gets 5x Pro. Limits are enforced per
  API key. Detailed per-endpoint throttling is documented inside the LlamaCloud product after sign-in.
sources:
  - https://www.llamaindex.ai/pricing
  - https://developers.llamaindex.ai/python/cloud/
notes: Public pricing page mentions rate-limit scaling between tiers but does not publish concrete
  RPS/RPM numbers. Replace with per-endpoint values when LlamaCloud publishes a rate-limits doc.
responseCodes:
  throttled: 429
  quotaExceeded: 429
limits:
  - name: API requests (Free)
    scope: api-key
    metric: varies
    limit: 'see plan tier; lowest tier in LlamaCloud'
  - name: API requests (Starter)
    scope: api-key
    metric: varies
    limit: 'see plan tier'
  - name: API requests (Pro)
    scope: api-key
    metric: varies
    limit: 'see plan tier'
  - name: API requests (Enterprise)
    scope: contract
    metric: varies
    limit: '5x Pro rate limits'
policies:
  - name: Tiered scaling
    description: Per-key request limits scale by plan tier; Enterprise contracts receive 5x Pro rate limits.
  - name: Backoff
    description: On 429, clients should retry with exponential backoff and respect Retry-After when present.
  - name: Credits as soft quota
    description: Per-month credit allotments act as a soft quota; PAYG continues at $1.25/1,000 credits up to a per-tier monthly cap.