Mantella is a Skyrim and Fallout 4 mod which allows you to naturally speak to NPCs using Whisper (speech-to-text), LLMs (text generation), and xVASynth / XTTS (text-to-speech).
This allows specifying generating only subtitles rather than text-to-speech, so long as the facefx_folder is set (automatically or otherwise; if the xVASynth or XTTS folders for FaceFXWrapper exist, then those will be used). It is annoying that we still need FaceFXWrapper even for silent audio files, but it's ultimately not too much of a price, considering I can't find documentation on the .lip format and this avoids risking any crashes due to missing files.
This allows specifying generating only subtitles rather than text-to-speech, so long as the
facefx_folderis set (automatically or otherwise; if the xVASynth or XTTS folders for FaceFXWrapper exist, then those will be used). It is annoying that we still need FaceFXWrapper even for silent audio files, but it's ultimately not too much of a price, considering I can't find documentation on the .lip format and this avoids risking any crashes due to missing files.