Agent

AgentPool

A unified agent orchestration hub that lets you configure and manage heterogeneous AI agents.

View Documentation
Multi-agent orchestration
YAML-based configuration
MCP bridge support
Docker and sandbox environments

Getting Started

Learn how to get quickly setup using AgentPool with ACP.

1

Install AgentPool

Run uvx --python 3.13 agentpool@latest serve-acp --help to get started without local installation.

2

Create a configuration file

Define your agents in a YAML configuration file specifying roles, models, and tools.

3

Configure your editor

Add AgentPool to your editor's agent_servers settings with the uvx command and your config file path.

4

Start coding

Select AgentPool from the available agents in the agent panel to start a new thread.

Learn more about ACP

The Agent Client Protocol enables editors to support any ACP-compatible agent without custom integrations.