Hello 
There's been a fair profundity of WiFi modems knocking about - e.g. WiFi232, WiModem etc. These are used for connecting to online BBS's or even running them (e.g. https://www.telnetbbsguide.com/bbs/a-80s-apple-ii-bbs/). They all use the ESP8266 module and some form of RS232 level shifter. None of them are open source 
even if they use some open source firmware.
I've just released some open source hardware which works with open source firmware:
https://github.com/stardot/esp8266_modem/releases
It implements all of the RS232 signals (DSR/DTR/DCD/RING/RTC/CTS/RX/TX) and, being open source, anyone can make them (and sell them if they so desire).
The project is available here: https://github.com/stardot/esp8266_modem/releases
Build instructions and BOM are in the Wiki. To programme it you just use the Arduino IDE and a cheap usb-serial adapter.
I've also been working on tweaking the Zimodem firmware (more fully featured) to work with the hardware here:
https://github.com/drdpj/Zimodem - currently it uses software implemented RTS/CTS, but I intend to adjust it to use the built in hardware flow control the ESP8266 provides.
I'm not intending to make these, but I'm happy to help anyone who wants to!
d.
Thank you for doing this!
I've got a slightly different WiFi modem in the works for the Apple II. Instead of a generic RS-232 device my design is an Apple II card. It also uses the ESP32 which is faster than the ESP8266 and communicates with it using i2C which is significantly faster than RS-232 style serial. Throughput should be significantly faster, probably pretty close to the theoretical maximum the Apple II bus can deliver, as that's probably the bottleneck. The down side will be that it will require new driver code because it does not use a Z8530, 6551 or 6850 serial chip like most Apple II serial devices do.
Edit -- I should say, close to the theoretical maximum that the Apple II bus can deliver w/o using DMA. But anyway, I'm expecting performance that should be easily 4+ times as fast as is possible using any normal Apple II serial port.
*nod* - that's a substantially different sort of beast. Will it be an open design?
This obviously is a generic modem, so is just designed to be plugged into anything that you could have plugged an old hayes serial modem into from TRS-80 to BBC Micro, to Apple II to Mac to PC to SGI O2
The real point wasn't to reinvent the wheel though, it was to liberate it.
I haven't decided on whether the design will be "open" or not yet. It sort of depends on how well a couple of other projects I'm working on for the Apple II are successful or not. In those cases I'm looking at actually selling assembled boards. It will be more or less at the break even level, as it does not appear that making products for vintage computers is a viable business, only a hobby. A few years ago when the software engineer job market was better I had a lot more ability to fund hobbies. Unfortunately in today's tough times I'm not able to subsidize things nearly as much.
Certainly my experience of making designs open has been that people still want you to make them, and pay you to do so. The benefit is that other people learn from what you've done and when you get sick of making them, if there's still demand, the design doesn't die, someone else motivated can pick it up and run with it
Win-win really!
Well, all of my designs are pretty simple and mostly old-school, so if anyone really wanted to reverse engineer them it wouldn't take much. If I get sick of making something I would probably release the info at that time. I'm working on learning FPGA development, that may be a little more involved for some things. If I can get over the hump on a few things as far as dealing with modern tools I should be able to make a lot more progress.