Another way to look at it: Actyx is a durable and reliable messaging middleware without any central brokers. You can publish events at any computer (e.g. RPi, Android Phone, Tablet) at any time — with perfect availability! — and receive these events at other computers if there is some network path to get them there. You can use streaming data queries [4] to listen to exactly those updates your app is interested in, or use our machine-runner library [5] to execute replicated state machines driven by these events without coordination overhead.
We built Actyx as we helped factories automate their manufacturing processes (see this J on the Beach talk [6] for more background); the factory environment is uniquely challenging in that it requires uncompromising reliability. The reason we’re open-sourcing this technology now is that we see it being valuable outside of this narrow field. You can use it to build resilient applications that cannot be stopped by an unavailability of AWS/Azure/GCP or your SaaS provider of choice — whatever the reason.
We would love to hear your feedback and see other use-cases profit from decentralised local-first software.
[1] https://github.com/Actyx/Actyx
[2] https://www.inkandswitch.com/local-first/
[3] https://www.local-first-cooperation.org/
[4] https://developer.actyx.com/docs/reference/aql