Cron-jobs, workflows, Discord bots, home automation, polling websites, handling webhooks, are perfect use-cases for Membrane.
No other platform gives you complete visibility into what your code is doing, has done, or can do.
Membrane gives you the power to easily and predictably interact with web services via their APIs and easily build useful abstractions and functionality.
Membrane is similar to an operating system; you deploy programs to it and it takes care of communication, logging, resources, etc.
In Membrane, you build custom functionality (bots, workflows, etc) by writing programs. Programs can reference each other's nodes to create powerful abstractions.
As a starting user, you'll first install drivers to talk to APIs you care about, then write programs using the nodes
exposed by those drivers to build something useful to you or your team. You can always use
fetch and/or build your own
drivers if you want.
Check out The Program Directory for many drivers and examples, anyone is welcome to contribute new drivers or improve existing ones.
More information in the programs section.
As mentioned above, each running program exposes a graph. The combination of these graphs is what we call the user's Membrane Graph.
The Membrane Graph is a powerful abstraction with many benefits:
The graph guarantees that programs (especially ones you didn't write) can do what they say they do, and nothing else. It also enables visibility into everything a program has done. Nothing is opaque in Membrane.
Membrane programs can't directly make network requests. Instead, they interact with
the world via graph nodes. Even
fetch uses the graph behind the scene.
Membrane programs are durable. This means you don't need to store data in a database or file to make it persist, the state of your program (the JS heap) is transparently and efficiently persisted every time it changes.
To keep data around, put it in the
state object and that's it.
Promises can be
await'ed indefinitely without worrying about execution timeouts.