README

Website
   • 
  Documentation
   • 
  X / Twitter
   • 
  Community Slack
   • 
  Forum





The Serverless Framework – Makes it easy to use AWS Lambda and other managed cloud services to build applications that auto-scale, cost nothing when idle, and result in radically low maintenance.

The Serverless Framework is a command-line tool with approachable YAML syntax to deploy both your code and cloud infrastructure needed to make tons of serverless application use-cases, like APIs, front-ends, data pipelines and scheduled tasks. It's a multi-language framework that supports Node.js, Typescript, Python, Go, Java, and more. It's also completely extensible via over 1,000 plugins which add more serverless use-cases and workflows to the Framework.

Actively maintained by Serverless Inc.



 Serverless Framework - V.4


  
    
  




October 2026 – V.4 continues to feature significant updates. Review them all below. Recent releases added Web Apps on AWS Lambda Web Functions, MCP server hosting on AWS Lambda, a rebuilt Dev Mode, built-in support for AI coding agents, Sandboxes (isolated, ephemeral compute on AWS Lambda), native Amazon Bedrock AgentCore support, Managed Instances, Durable Functions, and built-in AWS Login & SSO. As always, we are more excited about the serverless future than ever.

 New Features In V.4

Here's a list of everything that's new in V.4, so far:

- Web Apps – Run Express, Fastify, Hono, Next.js and other Node.js HTTP servers on AWS Lambda Web Functions. Define a server under the top-level webApps block and serverless deploy packages it, creates the web function and its HTTPS endpoint, and prints the URL. Your code listens on process.env.PORT as it does locally — no handler to export and no adapter to install. Custom domains are served through Amazon CloudFront. More info here.
- MCP Server Hosting – Deploy MCP servers written with the official MCP TypeScript SDK to AWS Lambda behind API Gateway with response streaming, via the mcp section in serverless.yml. More info here.
- AI Coding Agent Support – serverless agent setup installs the bundled Agent Skills that teach Claude Code, Codex, Cursor and other agents to work with the Framework, and the serverless agent commands give agents structured access to your service. When a command runs under a detected AI coding agent, the CLI never opens a prompt and shows progress as plain lines.
- Sandboxes – Deploy isolated, ephemeral compute environments on AWS Lambda — ideal for untrusted or per-session workloads such as AI agents and code execution. More info here.
- Amazon Bedrock AgentCore Support – Define AI agents, memory, tools, gateways, browsers, and code interpreters directly in serverless.yml via the ai property, and manage them with the serverless agent commands. More info here.
- AWS Login & SSO – Set up AWS credentials via browser-based flows with serverless login aws and serverless login aws sso.
- Deployment Diffs – Preview how a d

Top issues

Discussions

Latest release

4.46.0

 Features

- Run web apps in serverless dev. serverless dev now runs a service's web apps on your machine, in the same session as its functions and with no redeploy. Requests to the web app's real URL (and a custom domain in front of it) reach your local server, which gets the deployed environment variables and the web app's execution-role credentials. A local URL on localhost serves it with your framework's hot reload. Set webApps..dev.command to start a dev server such as next dev, or leave it out to run entry from source. When your machine doesn't answer in time, the deployed co