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: Prisma
providerId: prisma
generated: '2026-09-17'
method: derived
source: >-
  openapi/prisma-postgres-management-api-openapi.json (https://www.prisma.io/openapi.json),
  https://www.prisma.io/pricing
created: '2026-05-04'
modified: '2026-09-17'
supersedes: >-
  The 2026-05-04 bulk-sweep version, which was method: generated and cited Accelerate-era limits
  (the P6009 5MB threshold, Accelerate operation quotas) from products no longer in the line.
tags:
  - Database
  - Postgres
  - Rate Limiting
description: >-
  Prisma publishes NO numeric rate limits for the Management API. What it publishes instead is the
  runtime signal: 38 of 115 operations declare a 429 response, and Retry-After is the only
  response header modelled anywhere in the contract. Quantitative ceilings are commercial rather
  than throttle-based — each plan's included monthly operation and request quota, with overage
  billed per unit and a hard spend limit on by default.
limit_count: 0
limit_count_note: >-
  An honest zero. No published per-second, per-minute or per-hour limit exists for
  api.prisma.io. The entries below are the RUNTIME SIGNALS and the COMMERCIAL QUOTAS, which are
  real and measurable, but neither is a documented rate limit.
response_signals:
  status_on_exhaustion: 429
  operations_declaring_429: 38
  operations_total: 115
  headers_returned:
    - name: Retry-After
      modelled_in_contract: true
      note: The only response header declared anywhere in the published OpenAPI.
  headers_absent:
    - X-RateLimit-Limit
    - X-RateLimit-Remaining
    - X-RateLimit-Reset
    - RateLimit-Limit
    - RateLimit-Remaining
    - RateLimit-Reset
  agent_impact: >-
    An agent can back off CORRECTLY (Retry-After is authoritative) but cannot budget AHEAD —
    nothing tells it how much headroom remains before the next call is refused.
  variant_message: >-
    Compute operations return "Rate limit or Compute resource limit exceeded." on 429, folding a
    resource ceiling and a rate ceiling into one status with no way to tell them apart.
commercial_quotas:
  scope: per-workspace, monthly
  source: https://www.prisma.io/pricing
  behaviour_past_quota: >-
    On paid plans, usage past the included quota BILLS at the plan's overage rate rather than
    being throttled — until the hard spend limit (on by default) stops it. On Free there is no
    overage.
  quotas:
    - plan: Free
      postgres_operations_month: 200000
      compute_requests_month: 1000000
      storage: 500 MB
      databases: 50
    - plan: Starter
      postgres_operations_month: 1000000
      compute_requests_month: 5000000
      storage: 10 GB
      databases: 1000
    - plan: Pro
      postgres_operations_month: 10000000
      compute_requests_month: 20000000
      storage: 50 GB
      databases: 1000
    - plan: Business
      postgres_operations_month: 50000000
      compute_requests_month: 100000000
      storage: 100 GB
      databases: 1000
  metering_note: One database query = one operation; one HTTP request served = one Compute request.
payload_limits:
  - status: 413
    declared_on_operations: 20
    note: >-
      A 413 Payload Too Large is declared on 20 operations but no maximum body size is published
      anywhere in the contract or the docs.
policies:
  - name: Hard spend limit
    description: >-
      Every paid plan includes a hard spend limit, enabled by default, so usage-based pricing
      cannot produce a surprise bill. This is the real ceiling on an agent's spend.
  - name: Retry-After backoff
    description: On 429, honour the Retry-After header rather than a fixed backoff.
cross_reference:
  - conventions/prisma-conventions.yml
  - plans/prisma-plans-pricing.yml
  - errors/prisma-problem-types.yml
maintainers:
  - FN: Kin Lane
    email: kin@apievangelist.com