Unified Push Laravel Package

Product Decisions This Supports

• Build vs. buy vs. adopt for push delivery: The package models a single facade over APNS, GCM, and MPNS — useful as a reference for the shape of a push abstraction (per-platform drivers, one send call, token registration). That abstraction is worth reusing conceptually even if the code is not.
• Roadmap: mobile engagement surface: If push notifications are on the roadmap (transactional alerts, re-engagement, marketing), this tells you the capability is a commodity with a well-understood interface — scoping can proceed without vendor lock-in to this library.
• Legacy maintenance vs. migration: If a product already depends on this package, the decision it forces is migration. Last release 2016, zero dependents, 14 stars, and transports that are largely retired mean the real PM decision is "fund a replacement now" rather than "keep patching."
• Platform coverage decisions: MPNS is effectively dead (Windows Phone), GCM has been superseded by FCM, and the old APNS binary protocol has been replaced by HTTP/2. This is evidence for a decision to scope push to iOS + Android via current transports, and to defer desktop/web push to a separate track.
• Risk register input: Single-maintainer, no release in ~9 years, no dependents means no community safety net. This supports a "no new dependencies without a maintenance commitment" policy item.

When to Consider This Package

Consider it only in narrow circumstances:

• Legacy code you already own. An old Laravel app (roughly 5.x-era, PHP 5.6/7.0) that already uses the package and is not being actively developed. Reading it to understand existing behavior before a migration is legitimate.
• Architectural reference. As a worked example of driver-based notification dispatch, useful when designing an in-house service — treat it as documentation, not a dependency.
• Extremely constrained offline/internal tools where push volume is trivial and no current-SDK alternative is available for the target environment.

Look elsewhere — decisively — when:

• You need FCM (GCM is shut down), APNS over HTTP/2 with token-based auth, or anything shipped after ~2016.
• You need capabilities that did not exist in 2016: rich media, notification actions, silent/data messages, topic or segment targeting, delivery receipts, web push, VoIP push.
• You care about security posture: stale SDKs miss certificate/token rotation changes and platform auth deprecations.
• You need maintenance guarantees, SLA-backed support, or an actively patched dependency — none exist here (0 dependents, 1 release in 2016).

Realistic alternatives to evaluate: a managed provider (Firebase Cloud Messaging directly, AWS SNS/Pinpoint, OneSignal) or an actively maintained Laravel package with a per-platform driver design.

How to Pitch It (Stakeholders)

For executives: This is a small, MIT-licensed Laravel library that wraps three push providers — one of which (MPNS) is discontinued and another (GCM) has been replaced by Firebase. It has not been released since 2016, has no downstream dependents, and sits with a single maintainer. It is not a viable foundation for a new push capability, and adopting it would create a security and reliability liability we would have to fund later anyway. The strategic takeaway is positive: the design pattern (one interface, swappable platform drivers) is sound, and the underlying capability is commoditized. We should build on a supported provider and own only the thin dispatch layer.

For engineering: Reuse the interface idea, not the code. Expect a migration if anything internal already depends on it — the transports are outdated and the package predates modern auth flows. Recommended path: define our own small PushDriver contract, back it with FCM for Android and APNS HTTP/2 for iOS, and keep the provider behind an adapter so we can swap vendors without touching product code. Budget should reflect a fresh implementation, not an integration of this package.

Recommendation: Do not adopt. Fund a thin in-house abstraction over a current, maintained provider; migrate any existing usage off this package as part of normal maintenance.