Parallel MCP agent

Using the Parallel MCP server to create an Agent that can search the web using Parallel’s AI-optimized search capabilities:

"""MCP Parallel Agent - Search for Parallel

This example shows how to create an agent that uses Parallel to search for information using the Parallel MCP server.

Run: `uv pip install anthropic mcp agno` to install the dependencies

Prerequisites:
- Set the environment variable "PARALLEL_API_KEY" with your Parallel API key.
- Set the environment variable "ANTHROPIC_API_KEY" with your Anthropic API key.
- You can get the Parallel API key from: https://platform.parallel.ai/
- You can get the Anthropic API key from: https://console.anthropic.com/

Usage:
  python cookbook/14_tools/mcp/parallel.py
"""

import asyncio
from os import getenv

from agno.agent import Agent
from agno.models.anthropic import Claude
from agno.tools.mcp import MCPTools
from agno.tools.mcp.params import StreamableHTTPClientParams
from agno.utils.pprint import apprint_run_response

server_params = StreamableHTTPClientParams(
    url="https://search-mcp.parallel.ai/mcp",
    headers={
        "authorization": f"Bearer {getenv('PARALLEL_API_KEY')}",
    },
)


async def run_agent(message: str) -> None:
    async with MCPTools(
        transport="streamable-http", server_params=server_params
    ) as parallel_mcp_server:
        agent = Agent(
            model=Claude(id="claude-sonnet-4-20250514"),
            tools=[parallel_mcp_server],
            markdown=True,
        )
        response_stream = await agent.arun(message)
        await apprint_run_response(response_stream)


if __name__ == "__main__":
    asyncio.run(run_agent("What is the weather in Tokyo?"))