Hi HN, I built MockNest Serverless because we had a very practical problem: our application needed to call APIs from a cloud development environment, but some of those APIs were not accessible from there. Even when test environments existed, setting up the right test data was difficult. Edge cases were especially painful, and for some government APIs it was pretty much impossible to create the exact situations we needed (they had a static data set).

So I created MockNest: a mock runtime for AWS Lambda that is compatible with WireMock mappings and can serve REST, SOAP, and GraphQL mocks. The idea is that you deploy it into your own AWS account and point your app or tests at it instead of the real external service.

Once we had more mocks, mock maintenance became the next problem. So I added AI assistance using Amazon Bedrock to help generate WireMock-compatible mappings from API specs. Its optional to use, so if you dont use ai endpoints - no usage is generated on bedrock.

It also supports persistent mocks in S3, API key or IAM auth, webhooks/callbacks, and response streaming/SSE simulation.

MockNest is open source, and it won the Creative Track Award in the AWS 10,000 AIdeas competition in April.

I would really appreciate some feedback on whether this would be useful in your own integration testing setup, and whether the README makes the use case clear.