seanhealy / lightphoneproxy Goto Github PK

Currently the preferred service (SMS or iMessage) is stored in memory. This should be moved to a database. Thinking SQLite would work nicely until there is a better client server break (#2) where moving to a more scaleable db would make more sense.

Likely would be good to use https://github.com/typeorm/typeorm or something like it to be safer with any db moves in the future.

The codebase is currently a bit spaghetti-ish. A bit of a refactor to better organize things would be prudent.

The getting started steps currently leave a lot to be desired. A better guide here would be nice.

In order to allow for #1 the service should be broken into a client and a server.

The Client would run on a user's Mac to communicate with the Messages app and the server would be responsible for managing proxy numbers, receiving web hooks, etc.

A web socket connection would likely be a good communication mechanism between clients and servers that would also remove the need for a tool like ngrok.

The biggest fixed cost of running this are the proxy numbers. These are also very shareable. Since it's only LightPhone users who will be sending messages to these proxy numbers we can identify the contact that they are proxying to by correlating the proxy number to the user who's light phone sent the message. Likewise we can do the opposite for the other message direction.

There would be other considerations when we get to here at any scale such as passing on the messaging costs but at a low scale that could be handled with donations or with some other ad-hoc mechanism.