Basic Gemini MCP client

1. AWS Lambda Code (API Gateway Integration): This is a custom MCP that provides three main tools:

a. get_distribution_info - Fetches CloudFront distribution details
b. fetch_logs - Retrieves CloudFront logs from CloudWatch
c. analyze_logs - Analyzes logs for issues and suggests remediations

The Lambda function has two main endpoints:

• /list_tools (GET): Returns available tools and their schemas
• /call_tool (POST): Executes the requested tool with given parameters

2. Client Code:

graph TD
    A[User Input] --> B[Fetch Tool Schemas]
    B --> C[Initialize Gemini AI]
    C --> D[Interactive Loop]
    D --> E[Call Tools via API]
    E --> F[Process AI Response]
    F --> G[Show Results/Get Follow-up]
    G --> D

Loading

Key Components:

1. API Integration:

async def call_tool(session, tool_name, parameters):
    # Makes HTTP POST requests to Lambda API

2. Gemini Integration:

client = genai.Client(api_key=GEMINI_API_KEY)

3. Conversation Flow:

• Gets CloudFront distribution ID from user
• Uses Gemini AI to determine which tools to call
• Makes API calls to Lambda function
• Processes and displays results
• Handles follow-up questions