Back to vendors
P

Pipecat

Also known as: pipecat-ai

Visit site
Entry pricePipecat Cloud from $0.01 per active agent minute. The open source framework is free.Full pricing detail

Open-source framework for building real-time voice and multimodal AI agents.

Pipecat is an open source Python framework for real time voice and multimodal AI agents, maintained by Daily with community contributors under a BSD 2-Clause license. A Pipecat agent is a pipeline of processors that stream audio, text and video frames, handling turn detection, interruptions and context so the builder can focus on what the agent does. Builders swap speech, language and vision services across more than 200 integrated providers, and connect users over WebRTC, WebSockets, PSTN and SIP telephony (Daily, Twilio, Telnyx, Plivo, Exotel) or WhatsApp calling, with client SDKs for web, mobile and C++.

Pipecat Flows structures a conversation as a graph of nodes with their own tools and transitions, and multiple LLM agents can hand off to each other. The LLM calls the builder's functions to reach outside systems. Pipecat Evals tests an agent with scripted and simulated conversations run against the same pipeline that ships, and OpenTelemetry tracing and built in metrics show each turn and the services it called. A UIWorker lets an agent read and act on the builder's own web app through the Pipecat client.

Daily sells Pipecat Cloud, managed hosting in regions in the US, Europe and India billed per agent minute, and Pipecat Enterprise, which runs the same control plane against agents in the customer's own Kubernetes cluster. Pipecat Cloud offers HIPAA controls with a BAA. Long term memory and knowledge retrieval come from partner services in Pipecat's catalog rather than from Pipecat itself.

Vendor details

Canonical URL

https://pipecat.ai

Category

Agent infrastructure

Subcategory

Voice & multimodal agent framework

Funding status

Open source project maintained by Daily (daily.co), which sells Pipecat Cloud and Pipecat Enterprise.

Company status

first party product

Use cases & customers

Primary use cases

Real-time voice assistants and conversational botsCustomer-support and receptionist voice agentsMulti-agent voice systems (handoff, parallel fan-out, distributed)AI companions, coaches, and meeting assistantsMultimodal (voice + video + image) interactive experiences

Target customers

DevelopersAI product teamsVoice AI builders

Deployment options

Self-hosted (Python)Pipecat Cloud (managed)Local

Integrations

Vendor neutral orchestration covers more than 200 integrated AI providers, including STT (AssemblyAI, Deepgram, Whisper and others), LLMs (OpenAI, Anthropic, Gemini, Groq and more) and TTS (Cartesia, ElevenLabs and others). Transports include Daily WebRTC, LiveKit, SmallWebRTC, Twilio/SIP (phone), WebSockets, and WhatsApp. Client SDKs cover JavaScript, React, React Native, iOS (Swift), Android (Kotlin), C++ and ESP32. Pipecat also includes Pipecat Flows for conversations structured as graphs of nodes, a CLI, the Voice UI Kit and the Whisker debugger.

In practice

Building a voice agent means wrestling with interruptions, turn-taking, and latency before you even get to behavior. Pipecat handles the hard real-time parts, so you focus on what the agent does.

You don't want to be locked to one vendor's speech-to-text or voice. Pipecat is vendor neutral, letting you assemble your own pipeline from more than 200 integrated providers of speech to text, language models, voices and vision.

Your voice agent needs to reach phones, the web, and embedded devices. Pipecat ships client SDKs across web, mobile, and even ESP32, so one agent connects from many surfaces.

Agentic Index coverage score

10.5 / 14 capabilities · 75%

Integrations & Tool Calling Full

Function calling lets the pipeline's LLM call functions the developer registers to reach external services and APIs during a conversation, with handlers written in the agent's own code, and Flows nodes carry their own functions and actions.

SourcePipecat, docs.pipecat.ai (Function Calling, Flows Functions)read 2026-09-27

Workflow Orchestration Full

Pipecat Flows structures a conversation as a graph of nodes, each with its own task, tools and transitions, written as declarative flow configs or in code, with state shared across nodes, so the builder configures the workflow. Multiple LLM agents each own their tools and context and hand off control through activation and deactivation, and the pipeline itself composes processors frame by frame.

SourcePipecat, docs.pipecat.ai (Pipecat Flows, Multiple LLM Agents, Agent Handoff)read 2026-09-27

Knowledge Grounding & RAG Partial

Knowledge retrieval integrations (Moss semantic search injected into the context, Keenable web search over MCP) and function calls to the customer's own stores bring knowledge into each turn from services the builder wires in. Pipecat itself maintains no index over the customer's content, any index belongs to the partner, and context is assembled per run.

SourcePipecat, docs.pipecat.ai (llms.txt: Knowledge Retrieval services, Context Management)read 2026-09-27

Human Oversight & Guardrails Partial

Daily PSTN supports transferring a call to a person, and Flows limits the LLM to the functions of the current node, a constraint the builder sets. No mechanism holds an agent action for a person's approval. The budget guardrail floe-guard in Pipecat's catalog is a third party service, not Pipecat's own.

SourcePipecat, docs.pipecat.ai (Daily PSTN, Pipecat Flows)read 2026-09-27

Security, Identity & Governance Partial

Pipecat Cloud uses email OTP two factor sign in, organizations whose members hold roles that are not listed, and REST API keys scoped to organizations and agents. Administrative API calls are logged for a year. HIPAA controls with a BAA and GDPR are stated, and Daily's trust center at trust.daily.co is available on request, but no SOC 2 or other certificate is named for Pipecat Cloud.

SourcePipecat, docs.pipecat.ai (Security Guide, HIPAA Compliance, Accounts and Organizations)read 2026-09-27

Observability & Auditability Full

OpenTelemetry tracing, built into the product, visualizes conversation turns and the services each turn called, built in metrics report time to first byte, processing time and LLM and TTS usage per service, and TRACE logging records every frame. Pipecat Cloud keeps agent and session logs readable in the CLI, dashboard and REST API.

SourcePipecat, docs.pipecat.ai (OpenTelemetry Tracing, Metrics, Pipecat Cloud Logging and Observability)read 2026-09-27

Memory & State Persistence Partial

Conversation state is held inside a session. Context aggregators build the conversation history the LLM reads, context summarization compresses older turns in long sessions, and Flows state management shares data across nodes. Long term memory across sessions comes from partner services in Pipecat's catalog (Mem0, Memcode, MemorySync, Synap), whose memory layers belong to those vendors, not Pipecat.

SourcePipecat, docs.pipecat.ai (Context Management, Context Summarization, State Management)read 2026-09-27

Deployment & Data Residency Full

Pipecat Cloud deploys agents to Daily hosted regions the customer picks (us-west, us-east, eu-central, ap-south) for latency and data residency. Pipecat Enterprise runs the same agents in a self hosted region on the customer's own Kubernetes cluster (AWS EKS and Oracle OKE reference architectures), with secrets that never leave it, and the BSD 2-Clause framework self hosts anywhere.

SourcePipecat, docs.pipecat.ai (Pipecat Cloud Regions, Pipecat Enterprise)read 2026-09-27

Prebuilt Agents, Templates & Packs Partial

Complete single agent and multi agent example apps, Flows examples, recipes and a Voice UI Kit are published for developers to clone and adapt. Customers assemble their own agents from these parts, and Pipecat names no packaged agents ready to adopt.

SourcePipecat, docs.pipecat.ai (Pipecat Examples, Recipes, Flows Examples)read 2026-09-27

Triggers & Channel Coverage Partial

Agents take calls over a broad set of channels, including Daily PSTN and SIP dial-in, Twilio, Telnyx, Plivo and Exotel WebSockets, WhatsApp Business Calling, WebRTC and WebSockets from web, mobile and embedded clients. Dial-out places a call when the customer's application posts to the agent's start endpoint, and Pipecat Cloud webhooks report events outward. There is no schedule, event or queue that starts an agent on its own.

SourcePipecat, docs.pipecat.ai (Daily PSTN Dial-out, Pipecat Cloud Telephony, WhatsApp Business Calling)read 2026-09-27

Model Flexibility & Routing Full

The builder chooses every model. Pipecat swaps speech, language and vision services across more than 200 integrated providers, usually with one line of code and the builder's own keys, and Pipecat Cloud bills model providers directly to the customer.

SourcePipecat, pipecat.ai (home) and docs.pipecat.ai (service integrations) and daily.co/pricing/pipecat-cloudread 2026-09-27

APIs, SDKs & MCP Extensibility Full

The Python framework is the SDK, with a server API reference. Client SDKs cover JavaScript, React, React Native, iOS, Android and C++, Pipecat Cloud has a REST API, a Python SDK and a CLI, and an MCP transport serves a Pipecat bot as an MCP server other agents call.

SourcePipecat, docs.pipecat.ai (llms.txt, Pipecat Cloud REST reference, MCP Transport)read 2026-09-27

Testing, Debugging & Optimization Full

A built in evaluation harness, Pipecat Evals, tests the builder's own agent. It runs scripted scenarios (user turns with expected phrases, function calls, latency budgets or judged checks) and simulated scenarios (an LLM caller with a persona, a goal and a success criterion) against the same pipeline that ships, with eval suites run concurrently from a manifest and a pass or fail signal for each run.

SourcePipecat, docs.pipecat.ai (Pipecat Evals, Eval Suites)read 2026-09-27

Browser & Computer Use Partial

A UIWorker reads the customer's web app through accessibility snapshots streamed by the Pipecat client and drives it back, scrolling, selecting, filling inputs and clicking, without a person driving each step. It works only inside an app that embeds the Pipecat client and implements the command channel, so it controls that app's page only and does not drive a browser or desktop elsewhere.

SourcePipecat, docs.pipecat.ai (Controlling the UI)read 2026-09-27

The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Product updates

·Agent capabilityPartially Verified

Expressive conversation arrives in alpha in Pipecat 1.15.0: the LLM writes expression tags into its reply and the voice service performs them instead of reading them aloud. The LLM learns which tags its voice service supports, with adapters for Cartesia, ElevenLabs, Fish Audio, MiniMax and others. A perception feature also tells the LLM how the user spoke each turn, and new Pipecat Tone models read the user's tone and tag plain replies.

Bears on: Agent capability

View source
·Agent capabilityVerified

Pipecat 1.13.0 lets client apps send files and images to a running bot through a new RTVI send-file message, and the bot adds them to the LLM context. A new FileResolver downloads file URLs only from publicly routable or allowed networks, refuses redirects and caps files at 50 MB. The same release adds, in alpha, a frontend model that keeps talking to the caller while a backend model works through longer tool chains.

Bears on: Security / enterprise

View source
·Workflow orchestrationVerified

Pipecat v1.4.0 added realtime_service_mode on LLMContextAggregatorPair, the on_user_turn_message_added event, and RealtimeServiceMetadataFrame, changing how realtime speech-to-speech services write context and expose turn behavior.

Bears on: Workflow orchestration

View source
View all 4 Pipecat updates →Tracked since Jun 2026 · Verified from public vendor sources

Pricing

Pipecat Cloud from $0.01 per active agent minute. The open source framework is free.

Usage, billed per active agent minute by profile, plus metered telephony, transfers and recording.

Free tier

Included quota

The open source framework has no usage limits when self hosted. Pipecat Cloud has unlimited concurrency and includes Daily WebRTC transport free for 1:1 voice sessions. Agents bill per active minute by profile (0.5 vCPU at $0.01, 1 vCPU at $0.02, 1.5 vCPU at $0.03), with reserved warm capacity at a lower rate. Telephony, transfers, recording and Krisp noise cancellation are metered.

What is public

The free open source framework, the Pipecat Cloud agent minute rates by profile, reserved rates, and transport and telephony rates are public. Pipecat Enterprise (self hosted regions) and bundled inference pricing are not, and are quoted by sales.

Billing mechanics

The framework is free to self host. Pipecat Cloud is pure usage billing per active agent minute, at $0.01, $0.02 or $0.03 by agent profile, with optional reserved instances at a lower per minute rate to remove cold starts, plus metered telephony (SIP and PSTN) and a per event call transfer fee. Model, speech to text, and text to speech usage is billed by the underlying providers unless bundled at enterprise scale.

Cost watchouts

Model, speech to text and text to speech costs bill through the providers and are usually the largest line item. Reserved agents bill while idle, telephony and recording are metered, and a SIP REFER transfer costs $0.20 an event while other transfers keep minutes accruing.

Variable cost rationale

Pipecat Cloud bills purely by running agent minutes, and voice workloads run continuously during a call, so cost tracks directly with call volume and duration. Telephony, recording, and transcription add metered charges, and model, speech to text, and text to speech usage is billed by the underlying providers, so total spend at scale is spread across several meters. Concurrency is unlimited, so no plan cap stops the meter. Self hosting the framework shifts this cost to the customer's own infrastructure and providers, and 1:1 voice sessions over Daily WebRTC carry no transport charge.

Additional watchouts

The headline $0.01 per agent minute covers hosting only, and model and speech provider costs are separate and typically larger. Reserved agents bill while idle. Pipecat Cloud states HIPAA controls with a BAA and GDPR, but names no SOC 2 or other certificate.

Overage / add-ons

There are no plan caps. Usage bills by active agent minutes per profile, reserved minutes, and metered telephony, transfer and recording add ons.

Sales call required

No, self serve available

Free / trial

Free and open source framework under the BSD 2-Clause license, self hosted at no cost. Pipecat Cloud is pay as you go, with Daily WebRTC voice free for 1:1 sessions.

Lowest paid plan

Pipecat Cloud agent-1x at $0.01 per active agent minute, pay as you go

Commercial notes

The model is developer first and vendor neutral. The framework is free and portable, and Pipecat Cloud adds managed scaling and telephony with no lock in, since the same code self hosts.

Key ambiguities

Real cost depends on the LLM, speech to text and text to speech providers a team chooses, which bill outside the agent minute rate, plus telephony and recording. Pipecat Enterprise pricing is not public.

Missing data

Pipecat Enterprise (self hosted regions) and bundled inference pricing, and exact telephony rates by region, are not fully public.

Agentic Index verified 2026-09-27

Alternatives to Pipecat

The closest documented capability profiles to Pipecat among agent infrastructure platforms tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.

  • Bernstein10.5 / 14Fuller documented coverage on Human Oversight & Guardrails and Triggers & Channel Coverage
  • Anchor Browser11.0 / 14Fuller documented coverage on Human Oversight & Guardrails and Security, Identity & Governance
  • Haystack12.0 / 14Fuller documented coverage on Knowledge Grounding & RAG and Human Oversight & Guardrails
  • Inngest10.0 / 14Fuller documented coverage on Security, Identity & Governance and Triggers & Channel Coverage
  • Pydantic AI12.0 / 14Fuller documented coverage on Human Oversight & Guardrails and Memory & State Persistence
  • SmythOS12.0 / 14Fuller documented coverage on Knowledge Grounding & RAG and Security, Identity & Governance

Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.