viking2917
viking2917 on Hacker News
Posted by
viking2917 on Wednesday July 01, 2026 @02:32PM
from the show-hn dept.
https://chromewebstore.google.com/detail/bookxlate-archiveorg-book/mfefkdeojibnbabedjonpgmbgbfplbhe
There are lots of OCR tools and translator tools, but using them on books is very laborious, involving downloading images, lots of copying, window-swapping, 3rd-party software, and general dreck. Often the OCR has to be chained to the tranlation, leading to more labor.I often want to read books in the Internet Archive in languages I don’t speak (e.g. old books in Latin). So I built Bookxlate, a...
I updated the app to allow manual add/delete of pages. Then tap the page and do New Image to upload per page. Thanks for the feedback!
Thanks! I have not tried anything other than English, not sure how good the LLMs are for that. Did you use Tesseract or Gemini?
Once the page structure is set up from the images (via the directory upload), you can upload new images for each page, but I didn't include an option to just create all the pages manually. It's a good idea. Going to add that...
coincidentally, posted this over on Show HN today. OCR workbench, AI OCR & editing tools for OCRing old / hard documents.
https://news.ycombinator.com/item?id=43976450. Tesseract works fine for modern text documents, but it fails badly on older docs (e.g. colonial american, etc)
Posted by
viking2917 on Tuesday May 13, 2025 @03:06PM
from the show-hn dept.
https://github.com/viking2917/ocr-workbench
OCR on old documents is hard. OCR Workbench uses AI for OCR and provides an editing environment to clean things up, as is inevitably required.Inspired by this Hacker News post: https://news.ycombinator.com/item?id=43048698Backstory: I was having trouble producing transcriptions of Colonial American documents, which have their own unique challenges for OCR, and things like Tesseract fail...
Those of us "of a certain age" are grateful not to look at a screen of 30 videos, all of them illegible because our eyes are old :)
Depends a bit on what you want to accomplish.
I use Jekyll/Minimal Mistakes/Github pages for product blogs, where I just want to broadcast information or have it available for people to find through search, and where I need to own and brand the domain name itself.
I use Substack for a few projects where I want a stronger relationship with the audience - commenting, automatic email list building and distribution, and easy/attractive formatting. I just keep the Substack free, and don't worry about the domain name (x.substack.com) is fine.
Same. The view changing every time I click is very disorienting. and have to zoom back out after every click
Love this concept!
Implementation seems a bit erratic, e.g. tried it on the Richard I page (Richard the Lionheart), https://en.wikipedia.org/wiki/Richard_I_of_England
and get most random modern dates of research articles, and virtually none of the basic dates of his life (birth death, etc).
Book Roulette is a Chrome Extension that shows you a new, hopefully interesting, book each time you open a tab. It also shows you a few recent articles from around the web that mention the book.
There are a few other Chrome extensions out there that do this. 100 Million Books, for instance, is very nice. The reason I don't care for it is the books are completely random, and often super esoteric.
Book Roulette uses the backed from the book recommendation engine The Hawaii Project which scans over a thousand book review websites to see what people are talking about. Book Roulette pulls from the more popular books on a given day so the results are likely to be topical some way.
Enjoy! (It's free, but the Amazon and Bookshop links are wired to my affiliate account so if you buy I do get a commission).
Posted by
viking2917 on Friday August 27, 2021 @02:27PM
from the show-hn dept.
I run a social reading app call Bookship (think, virtual book club).
https://www.bookshipapp.comApp Tier on iOS, Android, Web, from a single source built with: Ionic V1 / Javascript
in-app Live Video Chat through JITSI
Deep Linking to app by Branch
Web hosting: EC2/Route53/Cloudfront
Web Analytics: Mixpanel
Server (old school: Mysql/PHP, MYSQL on AWS RDS server, PHP on the web servers @ EC2)
Source code mgmt: Github
IDE: Vs Code/Xemacs
Node + SimplePiE for ingesting and processing RSS feeds mentioning books.
Lifecycle email services through Mailgun.
Lifecycle text messages through Twilio
Blog via AWS Lightsail/Wordpress
Social media automation: Buffer
This is pretty cool, actually.
Bookship is social reading app for sharing your reading experiences with your family, friends and co-workers. Invite people to read along with you, whether they’re reading via a physical book, an ebook, even an audiobook. Post and react to thoughts, photos/videos, quotes, links and questions, in an easy-to-use chat-style interface. Get notified when others post and share your place in the book with them to stay in sync. Dog-ear passages with a quick photo on your phone, and even have Bookship extract the text from the page you took a picture of.
Thanks! Just launched and would love feedback. iOS and Android. Built with Ionic.
Posted by
viking2917 on Thursday August 03, 2017 @10:16AM
from the show-hn dept.
Posted by
viking2917 on Monday November 28, 2016 @04:33PM
from the show-hn dept.
Posted by
viking2917 on Wednesday January 13, 2016 @11:57AM
from the show-hn dept.
This was built at the CODEX books hackathon this past weekend at MIT.
Posted by
viking2917 on Wednesday January 13, 2016 @11:35AM
from the show-hn dept.
which books are we missing? want to take a look....thx.
Thanks. I can reproduce when I set "block sites from setting any data", caused by a 3rd party tool I use. I will have to figure some way to fix it. Thanks for reporting. In the meantime if you want to explore you'll have to turn that setting off... :(
Whoops. I use a 3rd party tool to manage user data and that requires access to local storage. I'll have to look into engineering that requirement out unless you have a user account. Sorry for the inconvenience.
Thank you this is great feedback. I agree with you the navbar is kind of a mess. I've got a code branch in progress to clean that up substantially.
I'm a recovering executive remembering how to code, and my design skills are awful. I have a designer friend giving me styling advice. It is too heavy. I was going for that hawaii feel but I think it's too much :)
My articles is getting a revamp shortly - the gist is, just as you can get book recommendations, you can get article recommendations about books - kind of a Flipboard for Books or StumbleUpon for Books. But that feature is still formative....i.e. not really ready for primetime.
Thanks for this. Keep it coming. You can find my email address on the About Us page if you want to send more.
Let me investigate. Neither of those should be required but perhaps something's busted. Browser you're using?
Thanks. Looks like jquery's not loading. either it's because of load, or perhaps you have javascript blocked? (more likely somehow load related. I'll take a look)
thanks for notes on design. I suspect the page loading is because my server's getting hammered right now :)
Very much so, like Hype Machine (which I love), with the caveat that THP is personalized, whereas Hype Machine is more one-size-fits-all in terms of the music it recommends, it doesn't really know whether I like EDM or Jazz and act accordingly.
It's around 800 right now - I curate all the source sites by hand at present, to ensure quality. If there's something you think I should add let me know...