- Joined
- Sep 21, 2025
- Messages
- 289
- Reaction score
- 680
- Awards
- 5
I've been playing around with a self-hosted LLM for copyediting my own work at home (thus not giving my info to machines to gobble up), and it occurred to me that adding a translation option was an easy addition.
What I have set up on a home server is a script that pulls a document apart, does the proofreading/copyeding/translation in 1000-word batches, and stitches it all back together. I set larger documents to go before bed, and wake up to the work done.
Since we're an English-only forum, I wanted to ask if anyone had any occult books that, as of yet, remain untranslated so I can give it a try and see how things go at scale with something other than my first test, Le Petit Prince. (I checked the rules, and this request doesn't seem to violate the rules as best as I could tell, but if it belongs somewhere else, please feel free to move it.)
What I need in the way of books that might work:
Please no suggestions with:
This is a test, so no promises as to outcomes or which books I pick. But I figured that if anyone could benefit from a translation that only costs electricity, it might be us.
What I have set up on a home server is a script that pulls a document apart, does the proofreading/copyeding/translation in 1000-word batches, and stitches it all back together. I set larger documents to go before bed, and wake up to the work done.
Since we're an English-only forum, I wanted to ask if anyone had any occult books that, as of yet, remain untranslated so I can give it a try and see how things go at scale with something other than my first test, Le Petit Prince. (I checked the rules, and this request doesn't seem to violate the rules as best as I could tell, but if it belongs somewhere else, please feel free to move it.)
What I need in the way of books that might work:
- PDF, DOCX, or EPUB format
- Text as text characters - no scanned pages unless it's scanned with text also present/OCR already done.
- Not rely on a lot of images (images get stripped out)
- Be in a major publication language (i.e. French, Russian, Japanese, Chinese, German, Spanish, etc.)
- Can be found easily online (Via Anna's or a DDG search)
Please no suggestions with:
- Obscure, hard to find books, or random things scanned on Archive.org
- Scanned pages with no OCR
- Languages that aren't very very widely spoken (i.e. Bambara, Chewa, Quechua, etc.)
- Books with a ton of images
This is a test, so no promises as to outcomes or which books I pick. But I figured that if anyone could benefit from a translation that only costs electricity, it might be us.
