Google (Gemini)

Configure Google Gemini to access b.well's healthcare data and services through b.well Health SDK for AI. This guide includes configuration parameters and code examples for quickly setting up the MCP client in Google's Generative AI API.

Prerequisites

Install the Google GenAI Python libraries:

pip install google-genai fastmcp

Required Credentials:

Credential	Source	Notes
Google Gemini API Key	Google AI Studio	Required for all requests
b.well User Token	OAuth 2.0 Token Exchange	Bearer token for user authenticatio

MCP Configuration Parameters:

Google Gemini connects to b.well's MCP server through the FastMCP client using Bearer token authentication. Configure the MCP client with these parameters:

Parameter	Value	Description
`mcp_fhir_agent_url`	`https://mcpfhiragent.{ENVIRONMENT}.icanbwell.com`	b.well MCP server endpoint (`client-sandbox`)
`auth`	`BearerAuth(user_token)`	Bearer token authentication wrapper
`model`	`gemini-2.0-flash`	Google Gemini model version
`tools`	`[mcp_client.session]`	MCP client session for tool acces

Code Example: Bearer Token Authentication

Use this approach when you have a user access token from b.well's OAuth 2.0 token exchange

import asyncio

from fastmcp import Client
from fastmcp.client.auth import BearerAuth
from google import genai

person_id = "{insert person id here}"
user_token = "{insert user token here}"
mcp_fhir_agent_url = "https://mcpfhiragent.client-sandbox.icanbwell.com"

mcp_client = Client(
        mcp_fhir_agent_url,
       auth=BearerAuth(user_token),
)

gemini_client = genai.Client(api_key="{insert your Gemini Key here}")

async def main():
    async with mcp_client:
        response = await gemini_client.aio.models.generate_content(
            model="gemini-2.0-flash",
            contents=f"Get Active Medications for person_id={person_id}",
            config=genai.types.GenerateContentConfig(
                temperature=0,
                tools=[mcp_client.session]
            ),
        )
        print(response.text)


if __name__ == "__main__":
    asyncio.run(main())