Introducing the Serverless MCP

‍

The new Serverless MCP allows you to debug issues with AWS serverless resources (Lambda, DynamoDB, IAM, API Gateway, etc.) within AI-Assisted IDEs like Cursor and Windsurf.

The MCP works by intelligently scanning your project's Infrastructure as Code files to understand the resources involved, then interacts with your AWS account to retrieve state, configuration, logs, metrics, and more. It brings back only the relevant information directly into your AI assistant, which also has access to your code. With all these elements combined, the AI has everything it needs to debug issues - no AWS console visits necessary.

The MCP is included in the Serverless Framework CLI, so it's easy to install. Check out the documentation to get started.

The Serverless MCP is one of many new features of Serverless Framework V.4. If you have upgrading or pricing questions, meet with us and we can help.

Key Features:

• Service-wide summary: Generate comprehensive overviews of your service, essential resources, configurations, deployment dates, invocation counts, and error reports
• Intelligent resource discovery: Automatically understands your project files and determines which AWS resources need review
• Error analysis and grouping: Identifies patterns in error logs to pinpoint recurring issues
• Up-to-date documentation access: Query current documentation for Serverless Framework and Serverless Container Framework

Powerful Tools at Your Fingertips:

• list-projects: Discovers all serverless projects in your workspace
• list-resources: Identifies all resources defined in your infrastructure configuration
• service-summary: Provides a consolidated view of your entire service in one call
• deployment-history: Retrieves chronological deployment events for your infrastructure
• aws-lambda-info: Diagnoses Lambda function performance issues and configuration
• aws-iam-info: Retrieves detailed information about IAM roles and policies
• aws-errors-info: Identifies and groups similar error patterns across your services
• docs: Accesses comprehensive up-to-date documentation
• and many more...

Getting Started

The Serverless MCP is built directly into the Serverless Framework CLI, making it immediately available if you're already using Serverless Framework. It required Serverless Framework version 4.13.0 or later.

Install or update Serverless Framework via

npm i serverless -g

Serverless MCP integrates easily with AI-assisted IDEs like Cursor and Windsurf through the Model Context Protocol (MCP). Setup takes just minutes using either stdio or SSE transport methods:

For Cursor:
• Open Settings > MCP
• Add a new global MCP server with the configuration:

{  
  "mcpServers": {    
    "serverless": {      
      "command": "serverless",      
      "args": ["mcp"]    
    } 
  }
}

‍

For Windsurf:
• Navigate to Settings > General > Cascade section
• Add a custom server with the same configuration as above

For SSE transport in either IDE, simply run `serverless mcp --transport sse` first, then point your IDE to the server URL.

For complete setup instructions and to learn more about what Serverless MCP can do for your development workflow, visit our documentation.

Experience the power of debugging your serverless applications without context switching - try Serverless MCP today.

‍