Since you seem to be on a implementation spree @lnkr we could probably use some inline translation like Phanpy does. They way its done there is pretty perfect, and when I am playing in Plamenu's interface I miss it right away

@wakest I do have it! Currently it supports LibreTranslate, DeepL and local LLMs. I had it running with some small local model on the same server for a while, on paper it's cool privacy-friendly setup but it still was too slow and resource hungry so I recently turned it off and haven't connected to anything else yet.
Phanpy uses Google Translate through some unofficial ways https://github.com/cheeaun/translang-api . It's quite effective but I'm not super happy about the idea of sending fedi posts to google while potentially breaking their ToS and don't want to risk getting any abuse countermeasures on my servers. I'll figure something out soon and I'll let you know

@lnkr I didn't realize phanpy was using google translate, interesting. so the libretranslate support is already working its just turned off atm?

@wakest libretranslate would be a separate thing to self-host, I tried it locally but wasn't happy with the resource usage and translation quality either, I don't currently have it. Maybe I can use some public libretranslate instance