Multi-Modal Orchestration
Coordinate voice, video, text, and images in one pipeline, with frame-level control over every step of the conversation.
Pipecat is the most widely used open source ecosystem for building voice and multimodal AI agents
Maintained by Daily with the support of the Pipecat developer community
Enterprises building with Pipecat
Production-grade agents run on Pipecat
Build, test and scale voice and video AI applications
Use Any Model
Swap speech, language, and vision services from any of 200+ integrated providers and services — usually one line of code.
Built for Real-Time
Frames stream through the pipeline as they're produced. No stage waits on the one before it, so the agent starts responding while you're still talking.
Human-like Conversation
Interruption handling, turn detection, and smart context management make agents feel natural to talk to.
Multi-Agent Workflows
Hand off to subagents for long-running tools and complex tasks, and use Pipecat Flows when a conversation needs to follow a defined path.
Any Transport: WebRTC, SIP, PSTN
Client SDKs and UI Kit
Metrics, Traces, OpenTelemetry
Orchestration, and more
Compose a pipeline, connect a transport, and add tools. Then test it and watch it run.
Pipecat's CLI scaffolds a runnable project and wires up your coding agent. It registers the Pipecat Context Hub's MCP server so your agent codes using Pipecat's current APIs, not its training data.
$ pipecat init
Project directory [pipecat-bot]: flower-shop
✓ Wrote flower-shop/AGENTS.md
✓ Wrote flower-shop/CLAUDE.md
? How do you want to build?
❯ Build with a coding agent (recommended)
Scaffold a runnable bot now
✓ Wrote flower-shop/GETTING_STARTED.md
✓ Registered the Context Hub MCP server
✓ Built the Context Hub index (2m 47s)
Your project is ready.$ pipecat init
Project directory [pipecat-bot]: flower-shop
✓ Wrote flower-shop/AGENTS.md
✓ Wrote flower-shop/CLAUDE.md
? How do you want to build?
❯ Build with a coding agent (recommended)
Scaffold a runnable bot now
✓ Wrote flower-shop/GETTING_STARTED.md
✓ Registered the Context Hub MCP server
✓ Built the Context Hub index (2m 47s)
Your project is ready.Run your agents on any infrastructure
From your own servers to fully managed cloud — pick the deployment model that fits your team, and change it as you grow.
Self-hosted
Pipecat is open source and runs wherever Python runs. Deploy agents on your own infrastructure.
Pipecat Cloud
Ship to production without managing infrastructure. Pipecat Cloud runs, scales, and observes your agents for you.
Pipecat Enterprise
Meet security, compliance, and procurement needs with Pipecat Cloud deployed in your own environment, including AWS, Azure, and GCP, backed by enterprise support.
Enterprise support
Get expert support and FDE hours from the team that created Pipecat.
Work with the engineers behind Pipecat — on architecture reviews, latency tuning, custom deployments, and more — from your first prototype to enterprise scale.
Talk to usReady to build?
Start building with the Pipecat docs, MCP, and dozens of ready-to-run examples for voice and multimodal agents.