While I can't offer a solution, it's not that the rips are quiet. It's most likely it's multi-channel audio and your player/headphones is doing a bad job of downmixing it to stereo. It is probably easier to just make sure you are grabbing releases with 2.0 tracks from the beginning.
Yeah that's why I end up playing back video content with MPV (in my case using the jellyfin-mpv-shim). I've bound a couple of hotkeys to enable a custom 5.1->2.0 downmix, and another to enable dynaudnorm (basically dynamic compression) so that, when I'm not using headphones, I have a hope in hell of hearing the dialog without loud scenes waking up the neighbors.
Look up ReplayGain. It analyzes and then adds metadata about the peak gains of each file, to the file, without the need to re-encode anything. Foobar2k natively supports it. Hopefully Plex also has support for ReplayGain.
This is the first thing I thought of. ReplayGain has been how this has been done for decades. It doesn't need to reencode the file, which will reduce quality, and there are programs that will scan folders to calculate the gain for each file. Other options suggested that reencode seem unnecessary and inferior.
Is this for movies and tv shows? I fix it at the output level and leave my rips unchanged. Check your TV or AVR for a “night mode” or “dialogue enhancement” or use the “reduce loud sounds” system setting if you’re playing back an AppleTV.
If the issue is specifically quiet dialogue and loud action scenes, get a dedicated center channel and boost its volume by a few dB.
Tdarr is probably your best bet. Its main focus is video but it used ffmeg as the backend, so anything it supports is supported in Tdarr (theoretically)
This, I use tdarr with the "migz convert audio" and "downmix & dynamic range compression" plugins to make sure all my videos have stereo audio channels which gives me a far more consistent experience across my devices