SMS Mobile App

The phone in your pocket is already an SMS gateway.

It has a SIM, a carrier agreement and an allowance somebody is paying for whether it gets used or not. What it does not have is a way for one person to hand it a list and a message and walk away.

SMS Mobile App is that missing half.

An admin imports recipients from a CSV or from the built-in persons directory of 15,851 rows, then assigns a send batch to volunteer devices that have opted in. Each device sends through its own SIM.

That directory is a virtualised table, which is what a list of that size needs if the screen is to stay usable at all. Templates map fields into the send flow, so a message can carry a name without anybody editing them one at a time.

The plumbing is a plugin.

A custom in-tree Capacitor plugin called NativeSms wraps the Android send call, and a foreground service keeps a batch moving while the application is closed or the screen is locked. There is no SIP relay in the path and no per-send fee beyond what the carrier already charges.

It is deliberately send-only.

No inbox. No READ_SMS permission, ever.

That is a product decision and a store decision at the same time. The message-reading permissions are among the ones a review examines hardest, and an application that never asks for one never has to explain why it wanted it.

The batch cap is enforced where it cannot be argued with.

Ten devices per batch, written as a Firestore rules invariant rather than as a check in the admin screen, with atomic parent-batch counters underneath so that two devices finishing together cannot both write the same total.

A rule the client enforces is a rule the client can skip.

That matters more here than in most products, because the clients are handsets belonging to volunteers and running wherever those volunteers happen to be. The only enforcement worth having is the kind that lives on the other side of the network.

It runs on Firebase's free tier, with Cloudflare Workers doing the work that would otherwise want paid functions.

That is the constraint the whole product is shaped by. No paid service in the critical path, and no server that has to be kept alive. That is what makes it possible for one person to keep a set of things running at the same time rather than one after another.

It is documented properly.

A public documentation site structured on Diátaxis carries 41 pages and roughly 33,000 words, which is the difference between a tool somebody else can operate and a tool only I can.

What it cost is the phones.

This design trades a paid gateway for a set of real devices that have to be switched on, charged and willing. A gateway scales by paying more. This scales by finding more volunteers, which is a harder thing to arrange on a Tuesday afternoon and a much cheaper one to run for a year.

It is live at smsapp.aoneahsan.com, with an Android application on Google Play and the documentation at smsapp-docs.aoneahsan.com.

Send-only was the first decision. Everything else rests on it. An inbox would have doubled the permission surface and changed what the application is for, and neither of those is a trade worth making for a feature nobody asked for.