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: Pipecat
providerId: pipecat-ai
created: '2026-06-21'
modified: '2026-06-21'
reconciled: false
tags:
- AI
- Voice
- Multimodal
- Agents
- Realtime
- Framework
- Rate Limiting
- Quotas
- Throttling
description: >-
  The open-source Pipecat framework (the pipecat-ai Python library) imposes no
  rate limits of its own - throughput is bounded only by the developer's own
  infrastructure and by the rate limits of the third-party AI services (STT,
  LLM, TTS, vision) and transports wired into a pipeline. The Pipecat Cloud
  REST control API (https://api.pipecat.daily.co/v1) is the only Pipecat-owned
  surface that could enforce limits; Pipecat Cloud advertises unlimited agent
  concurrency. Specific control-API request limits are not published and are
  not reconciled in this artifact.
notes: >-
  Verify any Pipecat Cloud control-API limits in the Pipecat Cloud dashboard
  and REST reference on reconciliation. Realtime media throughput depends on
  the chosen transport and the limits of integrated STT/LLM/TTS vendors.
sources:
- https://docs.pipecat.ai/api-reference/pipecat-cloud/rest-reference
- https://docs.pipecat.ai/overview/cloud
- https://github.com/pipecat-ai/pipecat
responseCodes:
  throttled: 429
limits:
- name: Framework (pipecat-ai library)
  scope: self-hosted
  metric: requests
  limit: none imposed by framework
  notes: Throughput bounded by developer infrastructure and integrated service limits, not by Pipecat.
- name: Pipecat Cloud Agent Concurrency
  scope: account
  metric: concurrent_sessions
  limit: unlimited (advertised)
  notes: Pipecat Cloud markets unlimited agent concurrency; underlying compute is billed per active agent-minute.
- name: Pipecat Cloud Control API Requests
  scope: account
  metric: requests
  limit: see provider documentation
  notes: REST control-plane request limits are not published; verify in dashboard/REST reference.
- name: Third-Party Service Limits
  scope: provider
  metric: varies
  limit: see each provider
  notes: STT/LLM/TTS/vision vendors enforce their own RPM/TPM/quota limits passed through the pipeline.
policies:
- name: Downstream Provider Limits
  description: Effective limits are usually set by the integrated STT/LLM/TTS/vision and transport vendors, not by Pipecat.
- name: Backoff Strategy
  description: Clients of the Pipecat Cloud REST API should implement exponential backoff with jitter and honor Retry-After on 429 responses.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com