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: Convai
providerId: convai-com
created: '2026-05-25'
modified: '2026-05-25'
reconciled: false
tags:
  - AI
  - Conversational AI
  - Rate Limiting
description: |
  Rate-limit posture for the Convai REST, streaming transcription, and Live
  WebRTC APIs. Convai's primary throttling surface is interaction quotas
  enforced per plan tier (see plans/convai-plans-pricing.yml) rather than a
  globally published RPM/QPS table; per-account caps and concurrency limits
  are negotiated for Scale and Enterprise tiers. The records below capture the
  general algorithm shape (account-level quotas plus per-request fairness)
  and should be reconciled with the live docs before production use.
sources:
  - https://convai.com/pricing
  - https://docs.convai.com/api-docs/api-reference/core-api-reference
headers:
  retryAfter: retry-after
responseCodes:
  throttled: 429
  quotaExceeded: 429
algorithm: account-quota
spendLimits: []
limits:
  - tier: Free
    resource: Interaction API
    quota: limited monthly interactions
    concurrency: low
  - tier: Indie
    resource: Interaction API
    quota: higher monthly interactions
    concurrency: standard
  - tier: Professional
    resource: Interaction API
    quota: larger monthly interactions
    concurrency: standard
  - tier: Scale
    resource: Interaction API
    quota: production-grade interactions
    concurrency: elevated
  - tier: Enterprise
    resource: Interaction API
    quota: negotiated
    concurrency: dedicated
  - tier: All
    resource: Streaming Transcription (wss)
    notes: 16-bit PCM mono 16 kHz audio frames; one stream per session
  - tier: All
    resource: Live API (WebRTC, beta)
    notes: Per-session WebRTC PeerConnection; metrics and audio data channels