This MCP server demonstrates a Peppa Pig themed MCP implementation deployed as a Serverless/MCP server using AWS Lambda and API Gateway. The server itself is exposed via the Streamable HTTP transport, and clients interact with the server through the MCP interfaces defined by the Model Context Protocol. To get started you bootstrap the AWS infrastructure with Terraform, which provisions the Lambda function and API Gateway endpoint, then you point your MCP client at the generated endpoint. The repo includes a Node.js client (src/client.js) that can be used to list resources, read resources, and call tools defined by the server. Run the client after the endpoint is available to observe the responses, such as listing the available resources (e.g., timeslots) and using tools like ordering tickets.

Key workflows you can exercise include:

• Listing resources exposed by the MCP server (listResources).
• Reading a specific resource (readResource) to retrieve textual content such as available timeslots.
• Enumerating available tools (listTools) and invoking a tool (callTool) like ordering tickets with structured parameters. This demonstrates how the server integrates with Lambda-backed infrastructure while still conforming to the MCP transport and protocol.

Note that this server relies on the Terraform deployment to establish the endpoint, after which the client communicates with the Lambda-backed MCP server via the endpoint URL provided by Terraform.

Prerequisites:

• AWS CLI configured with appropriate permissions
• Terraform installed
• Node.js and npm installed

Installation steps:

1. Clone the repository: git clone <repository-url> && cd <repository-directory>

2. Install client dependencies (in the src directory): cd src npm install cd ..

3. Bootstrap and deploy the MCP server (Terraform): cd terraform terraform init terraform plan terraform apply

4. Retrieve the endpoint URL and set it as an environment variable used by the client/server: export PEPPA_MCP_SERVER_ENDPOINT=$(terraform output --raw endpoint_url)

   If you need, run from the repo root unsetting and re-exporting as needed

   cd ..

5. Run the client to interact with the MCP server: node src/client.js

Prerequisites recap: Ensure you have AWS credentials configured, Terraform installed and initialized, and Node.js with npm available to install dependencies and run the client.

Notes and tips:

• The server is described as an AWS Lambda-based MCP server accessed via API Gateway using the Streamable HTTP transport. The deployment is managed via Terraform and the client uses the endpoint exposed by API Gateway.
• The repository indicates it is not actively maintained; refer to the linked AWS samples repository for more examples of serverless MCP servers.
• After deploying, it may take a few seconds for the API Gateway endpoint to become fully operational. If you encounter connectivity issues, re-check the Terraform outputs and ensure the endpoint URL is correctly exported.
• The PEPPA_MCP_SERVER_ENDPOINT environment variable is essential for the client to locate the Lambda-backed MCP service. Keep this value secure as it points to your deployed endpoint.
• If you modify resources or tools in the Lambda-based server, you will need to re-deploy Terraform to reflect changes in the endpoint or configuration.