I personally went down the route of creating a mini wrapper around functionless/functions which allow me to run things locally via a server and then when it com

I personally went down the route of creating a mini wrapper around functionless/functions which allow me to run things locally via a server and then when it comes to deployment decide whether it should run via serverless or docker / on a vm via adaptors. It doesn't allow you to simulate a lambda function directly, but makes the assumption (that held so far) that all the different API's and runtimes usually come down to the same core principles (HTTP / Websocket / timed jobs). The core idea is exactly around testing which approach works best for lambdas / functionless. https://vramework.dev/ https://vramework.dev/ Disclaimer: The core maintainer and I haven't yet published the serverless adaptors. Also need to update the landing page, as the serverless/server switch is only one of the core principles.