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: Hyperscience
providerId: hyperscience
created: '2026-05-08'
# Provenance stamped 2026-08-11: this artifact was written by the API Evangelist
# bulk sweep dated 2026-05-08, not harvested from the provider. See roadmap#35.
method: generated
modified: '2026-05-08'
reconciled: false
tags:
- AI
- Document AI
- IDP
- Enterprise
- Automation
- Rate Limiting
- Quotas
- Throttling
description: >-
  Hyperscience runs as a per-tenant Hypercell deployment (cloud or on-prem),
  so rate limits and processing throughput are determined by the deployment
  sizing and contract SLA — there is no public managed-service rate-limit
  schedule. The platform claims 99.5% accuracy and supports
  high-throughput batch processing for enterprise customers.
notes: >-
  Numeric limits remain unreconciled until tenant deployment details are
  visible. The standard contract permits sustained throughput up to the
  provisioned compute capacity.
sources:
- https://www.hyperscience.ai/
- https://www.hyperscience.ai/products/api/
- https://www.hyperscience.ai/security/
responseCodes:
  throttled: 429
limits:
- name: Per-tenant deployment
  scope: tenant
  metric: requests
  limit: per provisioned compute capacity
  notes: Sized to contract throughput SLA.
- name: Document throughput SLA
  scope: tenant
  metric: documents/hour
  limit: contract-defined
  notes: Set by Hyperscience deployment plan.
policies:
- name: Backoff Strategy
  description: Honor Retry-After / 429 responses with exponential backoff.
- name: Capacity Planning
  description: Scale tenant compute via Hyperscience deployment options for higher sustained throughput.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com