What modes do you want to prioritize? I'm sure that you have a particular need but I don't agree that the stand-alone transmitter would be a widely sort after product, especially outside the experimenter or homebrew sector (where, in my experience, there may be more interest). More generally, with modern components and architectures it's easier than ever to make a transceiver, and it's not hard to arrive at a receive performance that exceeds that of an RSPdx by a good margin.
But there are always niches outside the mainstream. For example, I have an RSPdx and I sometimes use it in conjunction with a Zachtek stand-alone WSPR exciter for LF/MF operation. The Arduino-based exciter drives a homebrew Class D push-pull HexFET power amplifier, which contains the T/R switching needed to safely use the RSPdx receiver. The receive system runs on a W10 notebook, with SDRuno and virtual audio cables and serial port emulation to interface to WSJT-X.
For a less self-contained but much more flexible operation, I can also replace the Zachtek with the 1 mW transmit "drive" output from my TS-590S or TS-890S transceiver. However, that's only useful in a couple of test configurations since either transceiver blows the RSPdx out of the water in overall receive performance.
I realize that these examples are not the same use-case you're describing and, if you're interested in relatively low cost, fairly high performance, software defined transceivers you could also take a look at e.g. the Elad range. Otherwise, the contemporary stand-alone HF transmitter designs that I know of, and which match the performance class of the RSPdx, all sit in the experimenter/homebrew corner.
73, Peter.