agentbreeder-sdk

The official Python SDK for AgentBreeder — define, validate, and deploy AI agents programmatically.

Installation

pip install agentbreeder-sdk

For MCP server authoring support:

pip install "agentbreeder-sdk[mcp]"

Quick Start

from agenthub import Agent, Tool, Model, Memory

# Define an agent
agent = (
    Agent("customer-support", version="1.0.0", team="customer-success")
    .with_model(primary="claude-sonnet-4-6", fallback="gpt-4o")
    .with_prompt(system="You are a helpful customer support agent.")
    .with_tool(Tool.from_ref("tools/zendesk-mcp"))
    .with_tool(Tool.from_ref("tools/order-lookup"))
    .with_deploy(cloud="aws", region="us-east-1")
)

# Validate and export to agent.yaml
agent.validate()
agent.to_yaml("agent.yaml")

Multi-Agent Orchestration

from agenthub import Orchestration, KeywordRouter

pipeline = (
    Orchestration("support-router", strategy="router", team="customer-success")
    .add_agent("triage",  ref="agents/triage-agent")
    .add_agent("billing", ref="agents/billing-agent")
    .add_agent("returns", ref="agents/returns-agent")
    .with_route("triage", condition="billing",  target="billing")
    .with_route("triage", condition="return",   target="returns")
)

pipeline.to_yaml("orchestration.yaml")

Key Classes

All classes serialize to the same agent.yaml / orchestration.yaml format consumed by agentbreeder deploy.

Tier Mobility

The SDK is the Full Code tier of AgentBreeder. You can eject from a YAML config to SDK code at any time:

agentbreeder eject agent.yaml --output agent_sdk.py

TypeScript SDK

Looking for TypeScript / JavaScript? Install the official TypeScript SDK:

npm install @agentbreeder/sdk

See sdk/typescript/ for full documentation.

Links

• Documentation
• GitHub
• agent.yaml reference
• TypeScript SDK on npm