Overview
Google’s Agent Development Kit (ADK) is an open-source, code-first Python toolkit for building, evaluating, and deploying sophisticated AI agents. ADK includes native MCP support via the McpToolset class, enabling seamless integration with StackOne’s MCP server.
Official Documentation
Installation
Or with uv:
Quick Start
Connect to StackOne MCP and create an ADK agent:
import os
import base64
from google.adk.agents import LlmAgent
from google.adk.tools.mcp_tool import McpToolset, StreamableHTTPConnectionParams
# Configure StackOne account
STACKONE_ACCOUNT_ID = "<account_id>" # Your StackOne account ID
# Encode API key for Basic auth
auth_token = base64.b64encode(
f"{os.getenv('STACKONE_API_KEY')}:".encode()
).decode()
# Create agent with StackOne MCP tools
root_agent = LlmAgent(
model="gemini-2.0-flash-exp",
name="stackone_assistant",
instruction="You are a helpful assistant with access to data from connected platforms via StackOne.",
tools=[
McpToolset(
connection_params=StreamableHTTPConnectionParams(
url="https://api.stackone.com/mcp",
headers={
"Authorization": f"Basic {auth_token}",
"x-account-id": STACKONE_ACCOUNT_ID
}
),
tool_name_prefix="stackone_" # Tools become stackone_gong_crm_search_calls, etc.
)
],
)
# Run the agent
result = root_agent.run_sync("List recent Salesforce accounts")
print(result.content)
Environment Variables
STACKONE_API_KEY=<stackone_api_key>
GOOGLE_API_KEY=your_google_api_key # For Gemini models
Async Usage
For production applications, use async patterns:
import asyncio
async def main():
# Create agent (same as above)
root_agent = LlmAgent(
model="gemini-2.0-flash-exp",
name="stackone_assistant",
instruction="You are a helpful assistant with access to data from connected platforms.",
tools=[
McpToolset(
connection_params=StreamableHTTPConnectionParams(
url="https://api.stackone.com/mcp",
headers={
"Authorization": f"Basic {auth_token}",
"x-account-id": STACKONE_ACCOUNT_ID
}
),
tool_name_prefix="stackone_"
)
],
)
# Run async
result = await root_agent.run("Search recent calls in Gong")
print(result.content)
asyncio.run(main())
McpToolset(
connection_params=StreamableHTTPConnectionParams(
url="https://api.stackone.com/mcp",
headers={
"Authorization": f"Basic {auth_token}",
"x-account-id": STACKONE_ACCOUNT_ID,
"Content-Type": "application/json",
"Accept": "application/json,text/event-stream",
"MCP-Protocol-Version": "2025-06-18"
}
),
tool_filter=[
"salesforce_crm_list_accounts",
"salesforce_crm_get_account",
"gong_crm_search_calls"
] # Only expose specific tools
)
Class Name Update: In ADK v1.17.0+, use McpToolset (not MCPToolset). The old capitalization is deprecated but still works with a warning.
Resources
Next Steps
