Unlike the previous suggestion, my idea is this: once you’ve finished writing, the webpage gets published. However, when another user from a different country visits the page (which you can determine using Chrome’s language settings), GPT is run to translate the entire webpage. This only happens once. After that, users from that country will automatically see the translated text, making your blog readable to people from all over the world.
Background: A user of your website from China : )
I don’t think this fits into the purpose of Bearblog. This would probably be better achievd thorugh a browser extension or using an LLM-powered translator in another tab.
Server side verus client side, they are quite different.
There are multiple benifits for server side localization:
This is indeed a good idea!
I would love to see an integration of Deepl. They have a quite good privacy policy due to being located in Germany. My idea is to not dynamically translate posts but to integrate Deepls API into the post editor so the author can create multiple versions of their post in different languages.
The main Reason why this is so important is the incredible amount of overhead to maintain all the translated versions once your article or page collection grows.
And while articles would probably be just copy & paste with DeepL since their algorithms detect really well what is supposed to be translated and what is not ~ but even as a linux user who has an advanced browser with web-panels and a primary clipboard (select & paste with scroll wheel) it’s still an incredible amount of work if you have multiple sites.
And if the API of i.e. DeepL (which I highly reccommend) or LibreTranslate has costs attached, you can just “forward” that cost’s to those who want to use that feature.
PS: Vivaldi uses Lingvanex for their Integrated Translator ~ self-hosted on their servers ~ which works pretty well for translating snippets and webpages. They even made the process completely asynchronous so only what is actually seen get’s translated.
————
Well an I guess that’s the biggest CON against the idea of server-side translation: Most Webbrowsers already offer great inbuilt translation functionality ~ which can work completely ondemand.
I know Chrome Users are faced with Google Translate ~ and G-Translate really sucks. But even such a crappy tool is able to get the point across.
Most of the time…
————
And while server - side would certainly be most efficient ~ it’s still possible to implement ondemand translation with js inside the page. So it will detect the language and ask the user if it should translate the page.
————
After reconsideration of the above, I think the priority of this feature would mainly concern company sites.
But I do understand the importance to some. As I myself are planning a fully automated CI hook for my website which triggers the automatic translation of individual pages when the page visitor statistics hit a set amount of visits for a certain language.
CI = Continuous Integration = Pipelines which automate the full process from code that get’s updated to the finished product.