OpenClix

Rules live in a single JSON file rather than a dashboard or database. No schema migrations. No deploy window. You point the app at that config, either bundled in app resources or fetched over HTTPS, and the engagement behavior changes on the next launch. For teams used to waiting on infra tickets, that's the whole selling point.

OpenClix runs on-device, so there's no hosted control plane and no API key for local-first use. It's plain push notification logic and in-app messaging that the app runs itself. Notifications fire from the app, which keeps latency low. The catch: anything you'd normally do server-side, like remote segmentation, isn't part of the picture.

You bring the OpenClix client code into your repo as checked-in source instead of pulling it from a package registry at runtime. Read every line. Patch behavior directly. Skip the proprietary SDK lock-in. It also means you own upgrades, which is a fair trade if you value control over convenience.

The folder structure is legible. Interfaces and schemas are explicit. The repo ships examples, fixtures, and documented edit points. An AI agent can therefore read the config, adjust a rule, and hand you a reviewable diff. Ever tried explaining your internal conventions to a coding agent? Here you don't have to. Point it at the docs and it has enough structure to work without guessing.

OpenClix reacts to events your app already emits, so you don't build a separate tracking pipeline just to drive messages. You wire those events to rules once, and the same hooks cover onboarding, streaks, and re-engagement. The debuggable reasons for each trigger make it easier to see why a message did or didn't show.

Every rule decision surfaces a reason you can inspect. Why did that nudge fire at 2 a.m.? Why did another one never show? Instead of blind test sends against a live backend, you trace the logic locally. Small detail, big difference when you're tuning timing for a real user base.