generated: '2026-07-20'
method: searched
source: https://api-tools.memories.ai/visual-intelligence/getting-started/rate-limit
scope: per-account
notes: >-
  Rate limits apply per account; all API keys on an account share the same
  quota. Video-understanding model calls (Gemini, Nova, Qwen) follow the
  underlying provider limits rather than Memories.ai QPS. Enterprise customers
  can request higher custom limits.
limits:
- name: Standard APIs (upload, transcription, embeddings)
  limit_count: 1
  unit: QPS
- name: Scraping / task APIs (YouTube metadata via Rapid)
  limit_count: 12
  unit: QPH
- name: Scraping / task APIs (Apify channels)
  limit_count: 10
  unit: QPS
- name: Stream processing
  limit_count: null
  unit: concurrent streams
  note: Uses a max concurrent-stream limit (video + audio combined), not QPS.
retry:
  strategy: exponential-backoff
  initial_delay_seconds: 1
  max_delay_seconds: 32
  note: Doubling each retry. Repeated violations may temporarily suspend the API key.
response_429:
  envelope: '{"code":429,"msg":"Rate limit exceeded","data":null}'