I know what you mean, you want a calibrator like a
Fluke 5720A. Personally, I think they are just a smidgeon too expensive, even 2nd hand on ebay.
There is also
Project KX which looks like an excellent idea, but is not yet available.
A more old fashioned and realistically achievable way is to look up Conrads Mini Metrology Lab, then when you find the Kelvin Varley Divider is probably too much work, go for a Hamon Divider with /10 and /100.
Of course a decent calibrated 10V ref is a good idea. I've got a couple of LM399's as LTZ1000 territory is too costly. (LM399's and LT1001's etc I got for free from Linear as samples
)
It becomes a never ending project and beware - volt nuttery is involved.
Now the issue is the calibration procedure on your devices. Like you, I have meters that need an exact multiple of 1.9 on all ranges with a limited deviation. Other meters like the Keithley will allow 9-11 volts for it's 10V cal and let you program the actual voltage. This is great as my Chinese ref puts out 10.00275V so I have my 2 Keithleys exact.
My other meters are lower spec from 3.5 to 5.5 digit and so I can get away with using my PSU with my Keithley to measure the voltage output and I hit the cal button when I get exactly 190.000mV, 1.90000V, 19.00000 etc.
I need higher voltages and an accurate divider to do the rest.
Then there is resistance, and AC to deal with...