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: TrueFoundry
providerId: truefoundry
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:
  - AI Gateway
  - LLMOps
  - GenAI
  - Rate Limiting
description: TrueFoundry's AI Gateway exposes plan-level monthly request quotas (50k Developer, 1M Pro/Pro
  Plus, 10M+ Enterprise) but does not publish a dedicated per-second / per-minute throttling document
  on its public site. Per-key gateway rate-limit policies are configurable by tenants for downstream models;
  the platform-side request budget is enforced as a monthly quota.
sources:
  - https://www.truefoundry.com/pricing
notes: No public per-second or per-minute platform rate-limit numbers are documented. Quota caps are captured
  in plans/. Tenants can configure their own gateway-side rate limits per virtual key or model.
responseCodes:
  throttled: 429
limits:
  - name: Developer monthly request quota
    scope: account
    metric: requests_per_month
    limit: 50000
    timeFrame: month
    notes: Free Developer tier monthly cap.
  - name: Pro monthly included requests
    scope: account
    metric: requests_per_month
    limit: 1000000
    timeFrame: month
    notes: Included with the $499/month Pro subscription; overages billed.
  - name: Pro Plus monthly included requests
    scope: account
    metric: requests_per_month
    limit: 1000000
    timeFrame: month
    notes: Included with the $2,999/month Pro Plus subscription.
  - name: Enterprise monthly request floor
    scope: account
    metric: requests_per_month
    limit: 10000000
    timeFrame: month
    notes: Starting allotment; negotiated upward for enterprise contracts.
  - name: Pro Plus MCP tool-call quota
    scope: account
    metric: calls_per_month
    limit: 5000000
    timeFrame: month
    notes: MCP gateway tool-call quota included with Pro Plus.
  - name: Tenant-configurable gateway throttle
    scope: virtual-key
    metric: varies
    limit: tenant-defined
    notes: Customers can configure RPM / TPM limits per virtual key, model, or team in the AI Gateway
      control center.
policies:
  - name: Plan Quota Enforcement
    description: Monthly request quotas are enforced at the account level; overage on Pro is billed per
      vendor pricing, while Developer and Pro Plus appear capped at their published numbers.
  - name: Tenant Self-Service Throttling
    description: Tenants configure their own per-key, per-model, or per-team rate limits in the gateway
      control plane; these supplement the platform monthly quotas.
  - name: Standard 429 Retry
    description: Exceeding gateway-side limits returns HTTP 429; clients should implement exponential
      backoff with jitter and respect any Retry-After header.
maintainers:
  - FN: Kin Lane
    email: kin@apievangelist.com