Assuming water to be water that is less than 20% dissociated and setting the compressibility at zero we have a reasonable maximum at about 300Mbar with a temperature of ca. 1200K Redmer-Icaurus2011. Assuming $g=10m/s^2$ (close to Earth's) we can relate each 10 m height of the water-column to 1 bar pressure. 1 Mbar = 1000 bar for a water-column of 10 km height, makes 3,000km which is about the radius of Earth.
There must be a temperature of 12000 K at the center (maybe a small nuclear reactor like inside the Earth) and > 274 K at the surface if the atmosphere has a pressure of 1 bar.
This won't work exactly as described because I made a bit too many assumptions. E.g.: I ignored the gravitational pressure of the weight of the water-column completely which would add to the temperature as $(G * mass * mm)/(2 * k * r)$ with gravitational constant $G = 6.67428 * 10 ^{-11}$, Boltzman constant $k = 1.3806504 * 10 ^{-23}$, mm = molecular mass, and r = radius, which gives the temperature at the center by gravitation alone.
But it would allow for a couple of hundred kilometers of liquid water with a comfortable temperature, given a "well tuned" reactor in the middle and an atmosphere that holds some of the energy of the central star(s) back to keep the surface warm.
So: it's possible without any visible hand-waving and expensive unobtanium.