Skip to main content

Start a Project

Initialize a project, then build with a coding agent or scaffold a bot

Run Behavioral Evals

Test your agents with scripted scenarios and an LLM judge

Give Your Agent Context

Index current Pipecat docs and API source for your coding agent

Deploy to Cloud

Push your bots to production with one command

​
Requirements

  • Python 3.11 or later

​
Installation

The CLI ships with pipecat-ai as the optional cli extra. Install it globally with uv:
Verify installation:
All commands can use either pipecat or the shorter pc alias.
This gives you the built-in pipecat init and pipecat eval commands. The cloud and mcp commands each come from a separate package that you co-install with --with:
--with replaces the tool environment rather than adding to it, so every run must name every plugin you want. Installing with only one drops the other.
Commands from a plugin you haven’t installed still appear in pipecat --help, and print how to enable themselves when run.

​
Commands

  • pipecat init - Initialize and scaffold a new Pipecat project, the single entry point for new projects (built in)
  • pipecat eval - Run behavioral evals against your agents
  • pipecat mcp - Query the Pipecat Context Hub and register it with your coding agent (requires pipecat-ai-context-hub)
  • pipecat cloud - Deploy and manage bots on Pipecat Cloud (requires pipecatcloud)

​
Getting Help

View help for any command:

​
Next Steps

Build Your First Bot

Initialize your project and start building with pipecat init