Examples

Complete working examples for Pipecat Subagents.

All examples are available in the pipecat-subagents repository.

Local examples

These examples run all agents in a single process using the default AsyncQueueBus.

Single Agent

The simplest use case: one BaseAgent running a complete voice pipeline through the AgentRunner. No bus bridge, no multi-agent coordination.

Two LLM Agents

A greeter and a support agent that transfer control between each other. Demonstrates agent handoff with handoff_to() and the @tool decorator.

Two LLM Agents with TTS

Same as above, but each agent has its own TTS with a distinct voice. The main agent has no TTS — audio comes from the active LLM agent through the bus.

LLM + Flows Agent

An LLM router agent combined with a FlowsAgent for structured restaurant reservations. Demonstrates mixing agent types.

Parallel Debate

A moderator spawns three worker agents in parallel using task_group(). Each worker argues from a different perspective. Demonstrates task coordination.

Voice Code Assistant

A voice agent backed by a code worker that uses Claude Agent SDK with tools (Read, Bash, Glob, Grep) to answer coding questions. Demonstrates task-based agent integration.

Reference applications

These live in their own repos and exercise a coherent slice of the framework end-to-end, not just one feature.

Music Player (UI Agent)

A voice-driven music browser backed by a live Deezer catalog. Exercises the full UI Agent surface: voice/UI separation with attach_ui_bridge, <ui_state>-grounded Q&A, multi-turn deixis with keep_history=True, parallel fan-out via start_user_task_group with streaming worker results into a Discovery grid, ack-first ordering for slow tools, and a long-lived singleton catalog agent.

Distributed examples

These examples run agents across separate processes or machines, communicating via Redis or WebSocket.

Redis Handoff

The two-agent handoff split across separate processes using RedisBus. The main transport agent runs independently from the LLM agents.

Remote Proxy Assistant

A main agent connects to a remote LLM server over WebSocket using proxy agents. Demonstrates point-to-point distributed deployment.