Have you contacted the owner of the github repo? They seem to have at the very least put quite a bit of thought into the problem. No idea if they would be interested in doing so, but they might be able to arrange to find a small scale vendor to crank out a number of their design if they had a buyer lined up.
One of the things to keep in mind is that closed firmware on IoT gadgets can end up being a huge source of issues and frustration - although LoFence project may seem “unfinished” compared to something you can just order off an e-commerce site, so many of those gadgets come with huge buyer’s remorse when it becomes evident that the developers did not understand, or really care about, your application to the degree that you do, and leave you stuck suffering from entirely fixable problems on any of the sensing, radio communication, or power/battery fronts. Getting hardware that hosts software which can be improved not only by the original developers but also by the end user (or a consulting resource of their choice) based on lessons learned until it is right up against the limits of the technology may be worth a little extra time and money up front.
Of course that doesn’t mean you can’t pursue two paths - if there’s something ready to go, buying those would be a way to gain a quick sense of where the real challenges of the application are. And the possible need to setup gateway coverage would be the same for any LoRaWAN solution.
And then sometimes you can find a cheap gadget on an e-commerce site, recognize that it has a halfway decent hardware design with a known and reprogrammable MCU at the center, and be able to get both the ready availability at a low price in the near term and the longer range option of replacing the factory firmware with something more optimal.