Not exactly what you asked:

Is your plan to also reduce resolution and bitrate? If not:

• It’s going to reduce visual quality anyhow.
• Are you sure “changing codecs” would save you as much space as you think it would?

It depends where your material comes from. If it’s ripped from streaming services it’s (generally) already as compressed at as it can be regardless of codec. E.g. Amazon mostly still provides their streams in h264 (a 20 yo codec) that’s very high quality at very small filesize.

If you still want to do that: I don’t believe there is one automated solution that will save quality and space across the board. (If there is, I’d like to hear about it)

A scenario I’d envision for myself:

1. tinkering with ffmpeg for test results
2. grouping material according to these results
3. set up a separate ffmpeg-script for each group
4. wrap the whole thing into a systemd service and make sure it survives reboots, i.e. picks up where it left.