I am increasingly doing this with all of my data, by publishing it as YAML, and rendering a dynamic (static) API representation in JSON–all done with the same approach I’m using to publish my website(s). You can get at all of the data I use across my API research in a single API Evangelist developer portal, which just aggregates all of the JSON APIs I’ve published across my network almost 100 Github repositories, and supporting sites.
It is becoming a kind of poor man’s serverless. I’m going to keep polishing my approach. Get better with my responses, and my approach to reading and writing data schema to the data storage folder in each of my repositories, which can then be read statically using JSON APIs I’ve pushed from this data, using Liquid. It is a pretty scrappy approach to serverless, but done in a way that takes the servers out of the equation for me, offloading the front-end and back-end work for my network of sites to Github. I am not sure where I’m going with this. Sometimes I get better results from a more straightforward API implementation on my Amazon infrastructure, but I am finding some interesting use cases, and seeing another side-effect I am enjoying–it is making my serverless infrastructure forkable and usable by others.