Show HN: Glossarie – a new, immersive way to learn a language (glossarie.app)
The idea is to introduce vocabulary and grammar whilst you read eBooks in your own language. I've found that it is easier to remember vocabulary 'in context' and with regular repetition. Plus you don't have to carve out dedicated time for language learning. Other apps require you to build a habit around various exercises or ‘games’, whereas lots of people already read books.
From testing with early users so far it's proving effective for building a basic understanding of a language and quickly getting to the point where you can read and broadly understand text in the target language. It’s even better in combination with other apps that help with listening/speaking like Pimsleur.
There were lots of technical challenges making this. It turned out to be (reassuringly) hard to get accuracy to an acceptable level, requiring a rabbit-hole into machine translation. There was a lot of testing required to optimise the engine that chooses the translations to show and to reduce the friction when reading books. And the backend to support uploading books is a beast in itself. I’d love to share details if there is interest.
Roadmap
- Accuracy - 100% accuracy is the target, but at present there can be errors. Feedback from users will be important here so that accuracy issues can be generalised and solved at scale. Errors can be reported within the app - please do so if you spot anything!
- Dynamic difficulty - rather than have a progression of difficulty levels I’d prefer to introduce vocabulary and grammar automatically in response to user progress, balancing against the friction of seeing unfamiliar words. There’s a lot ‘under the hood’ to manage this today, but plenty of room to improve.
- More practice features - to reinforce vocabulary/grammar and support writing, listening and speaking.
- Better eBook support - improving the formatting of eBooks within the app and providing more methods for finding good books to read.
Use of AI
- LLMs provided a step change in accuracy and have enabled a feature that explains translations and grammar to the user
- vastly improving the utility versus a year ago.
- I believe apps like this, which use AI to enhance or scale functionality rather than simply acting as a wrapper over APIs, will be the major beneficiaries as LLMs improve.
Take a look, and let me know your thoughts or questions!
159 comments
[ 3.3 ms ] story [ 206 ms ] threadI will expand the app into more countries once I'm happy none of the books have copyright restrictions there. Feel free to message me with the country you reside in and I'll take a look soon.
If you can license a modern book that someone would actually choose to read on their own, I'd pay for it. Bonus if I can sort/browse the available books by Goodreads (or similar) score. Prismatext makes it tedious to discover that readers didn't care for their modern books.
That said, I'd gladly pay you/the site to handle that for me (by paying more than the book's retail price). Hopefully the translation would also be better than anything I imported.
(Two sibling replies linked to sites that sell technical non-fiction. That is a very hard way to learn :-) )
https://www.tembo.app
https://chrome.google.com/webstore/detail/readlang-web-reade...
If OPs project is using LLMs, it could definitely be much higher quality when it's swapping out more than a few words.
https://www.tembo.app
Although it would be even better (for me) if one were able to import their Anki decks and have your app figure out the level of "competency" for each word. This is the biggest gripe I have with adopting a new language learning app: Having to re-learn vocabulary that I am already 100% confident in.
Also, I would love to be notified when the Spanish version comes around - is there any way for that?
- a scrollbar and search for the Online Library would be helpful
- switching difficulty levels in the middle of reading could be helpful. Or if you keep that on a separate page, returning automatically to the last open position. (I was floating between beginner levels to find the right amount of challenge)
Are you using APIs that are unavailable on iOS 16 and under, or is it a matter of testing? My understanding was that about 25% of iPhone users aren’t on iOS 17 (myself included!) so it’s a fairly large demographic
Shameless plug: I’ve identified the same problem and built an app that shows a new word every minute on the Menu Bar so I can learn a new word while working: https://apps.apple.com/us/app/wunderbar-learn-language/id647...
Although a koreader integration would be great, there are tablets with both Android and iOS, as well as eReaders with the former.
https://github.com/FreeLanguageTools/vocabsieve/
If you claim that you privately had the idea 15 years ago, it's possible; you just need credible, and credibly dated evidence.
I tried the practice a bit, and the explanations (generated by ai I guess?) were very nice. I met a bit of an unfair situation in one question. The sentence started with "They" and the options were Ils and Elles. However, the sentence in English didn't hint towards a gender, and I failed the 1d2 and got what felt like a sarcastic explanation.
I would suggest tackling dynamic difficulty and algorithmic selection of what words to learn, when, and how often, and then let improving LLMs handle accuracy improvements.
Also because while I absolutely love the idea for seamless Hinglish style integration (as opposed to say a side bar which just told you what some words would be in a different language) it does mean that I'm no longer really reading the book, I'm reading the content but not the author. I don't personally read anything that I'd want to alter like that, but I can imagine for others it might limit its use to 'trashy novella read while travelling' or something.
Tldr the idea is brilliant, but for me too it needs to not be for eBooks.
I'll see how easy it is. I get palpitations thinking about developing on another new platform. Java and Swift were a challenge enough to learn!
[1] https://support.babbel.com/hc/en-gb/articles/13752043233170-...
> We're sorry, the requested URL was not found on this server.
For any English word that came from French, use the French cognate, and pronounce it in the French way. Or at least the latter.
Just playing with your app for a bit, and it's pretty cool! had a few questions though:
1. Wondering about the decision of using English books and translating pieces into other languages vs starting with (for example) a Spanish book, and translating the other way? Also, would something like this be a future thought of plan? Because currently I'm trying to read more popular books in my target language, rather than English books (right now, my toy app is just highlight arbitrary text -> send to azure translate). I tried to upload my book into your app in Spanish, but I guess it only works rn if the source is in English? Basically, a mode for even more immersion would be killer (Ala either full-target-languge mode or upload target language books).
2. The practice mode is pretty cool! I like this format of "complete the sentence". It looks like it's not based on book content at all, right? Would be cool to practice based on what I'm reading.
3. I'm reading on an e-reader, so I'd reeeeally like a no-animation/no-scroll mode. On an e-reader, the paginated page refreshing can help to reduce ghosting. Even better if there could be an e-reader mode that can flash the screen to further reduce ghosting issues on those devices.
1. I would love to get it to work all the way from a few translations in the target language to a full translation, with a sliding scale in between. 2. It's not connected to the book content. One idea I have is an optional quiz at the end of a session to reinforce new vocabulary/grammar seen. 3. I'll see if I can remove the animation when using the page ahead/back buttons on Android.
Could you make it work with any versions lower than 17.0?
I prefer to learn by reading in the target language and translating to English as I go along.
I definitely didn't see what I expected when opening a book for the first time -- I can already read or watch content in Italian. What I do today is pause (or stop reading) when I encounter a word I don't know.
What I expected when picking a level was definitely to see all Italian, though in retrospect I can imagine it's near impossible to do that without lots of paraphrasing.
But to me personally (much as I think this space needs more things, and that you OP are awesome for sharing it) that I'd not personally use something which wasn't entirely in my target language, as I find the way I've learned languages best so far to be similar to my current workflow, and over time I have to look up fewer and fewer words.
The limitation now is getting consistently high accuracy for whole sentences - but something I'll keep working on as the underlying technologies improve.
This app lies in a sweet spot where no ongoing API calls are required, everything is pre-calculated (at moderate expense!), but LLMs can scale some of the more 'human' work like explaining translations or checking accuracy. Albeit with the quirks and inconsistencies inherent with the current generation of models.