Show HN: I built this to talk Danish to my girlfriend – works with any language

This is an interesting idea.

But, besides not being able to use your site due to the errors mentioned by other posters, both the examples on your website give me pause:

The example input "Kan you hjelmpe mig {yesterday}?" reads to me as "Can you help me yesterday?", but that's just a nonsense sentence and an odd choice for an example. The word for "help" is also misspelled, but presumably that would get corrected.

And the suggestion of "Jeg vil gerne handle i morgen" for "Jeg vil gerne go shopping i morgen", instead reads to me as "I would like to act tomorrow". A more idiomatic translation would be "Jeg vil gerne købe ind i morgen".

Interesting use case. By coincidence I am also learning Danish quite intensively. What I do (and that tends to work for me) is the following:

- I try to actively write something and when I hit a wall, I try in other languages: the important in this part is to get the text flowing from me. Eg, "du could har hjælpe mig, når jeg called du" (trying to write in Danish, "you could have helped me when I called you")

- I then paste it on ChatGPT and ask for a C2-level correction with explanations on vocabulary and grammar, and translations into a few other languages. Eg, my prompt would be """ "du could har hjælpe mig, når jeg called du"

correct to c2-level and explain grammar and vocabulary. give related examples. then translate to German, English, French and Italian. """

This is incredibly helpful to do everyday, especially if you are also learning passively by reading/studying by yourself. By the way, for those curious, here's how ChatGPT would correct the sentence: "Du kunne have hjulpet mig, da jeg ringede til dig." Interestingly, it assumed I meant "called" as in, "telephoned" (not my original intention). Translating into other languages helps you spot and get a sensibility for such cases.

As an aside, for me the most challenging part of learning Danish is the pronunciation. It's beautiful, but it doesn't map out too easily to written words :)

It doesn't seem to work with "any language", as claimed. It cannot process Chinese writing (nor can it process pinyin). "Error processing your text", is the message. Though it's certainly possible the volume of traffic has overwhelmed things and it would work under normal circumstances?

This is really cool!

I've done a lot of research into LLM translation for my product[0], and I'm currently working on a deep translation service that provides reliably human-level translations.

I don't know what model you're using, but GPT-4.1 is probably the best for your use case - it's in the top few % for nearly every language, and has a low standard deviation, while also being relatively low latency and low cost.

[0] https://nuenki.app

First, I applaud you building something and encourage it.

Next, I want to say I don’t understand these llm “apps”. I tried this in ChatGPT with the minimal prompt (that you can keyboard shortcut) “fill in curly” and it gave me exactly the correct full sentence plus even better alternatives.

Why would I use a separate app?

In general why would I use any of these separate apps vs my included ChatGPT that I’ll have a button away with included unlimited subscription

Not to be nag but the example translation at the bottom of the page is not correct. Shopping is a word used in danish.[0] The translation with the word handle does not correctly convey the suggestion of going out to multiple stores looking at clothes etc. to buy. It more comes across as wanting to go buy groceries, or doing some kind of general commerce.

A translation to standard spoken danish would be jeg vil gerne shoppe i morgen.

[0] https://ordnet.dk/ddo/ordbog?query=shoppe%20

> works with any language

Does it work in Tojolab'al (Mayan language spoken by 70,000 in Chiapas)?

I get mildly irked by apps saying they work in "all/any" language. Obviously, it works in whatever languages ChatGPT "knows". Although for marketing, I understand the former is easier to say.

fwiw - ChapGPT acts like it knows these kinds of low-resource languages, but it's difficult to tell how much of it is just hallucinating. (I am trying to do work in this area).

Take an intensive course. That's what I did to learn Dutch with my Dutch then girlfriend (now wife and mother of our kids). Intense, focussed and I was fluent at a higher level than many natives within a year (which says more about 50% of the country than it does me, but I digress).

I entered the text that was present in the placeholder, but I got an error message: "Unable to process correction Error processing your text. Please try again."

Quick feedback, the idea sounds interesting to me and I wanted to learn more. Feature request: more examples, or cached examples that would still work if the server crashes.

Tried it with some Dutch, only getting "Unable to process correction" errors. Good idea but seems it need a bit more error handling.

Interesting idea. Also got errors.

That withstanding, surely having the user select what their target language is, would help with processing and accuracy? It's a one time lift for the user, and would ensure I'm definitely getting back my target language, not a dialect or totally different (norwegian/swedish come to mind immediately)

It seems that my GPT credits have been deplated, thus the error :) Sorry about that! I decided that if I do ran out, I will continue the development - showing that is has popularity!

In the meantime, I am at work, so I don't think I can add the rate limiting now... but thank you for everyone for the input!

Sounds cool!

(Why doesn't ChatGPT make something like this, essentially plugins, for their web interface? You make a script with macros and put that on a "Marketplace" and people can run it... would remove the need for setting up a whole nother website just to wrap around prompts)

In the example text hjelmpe is not a word in Danish.

A long time ago, before SEO spam, I would Google parts of what I had written phrase by phrase (with proper nouns removed or generalized) and if the number of results was under a certain threshold, I'd assume I was wrong.

Me and my wife made something very similar (https://nativi.sh/). I find it's a really good use case for LLMs to actually help you learn rather than doing everything for you.

Also, not sure if you're getting hugged to death but I'm getting this in the interface but not seeing any network failures.

``` Unable to process correction

Error processing your text. Please try again. ```

I did a similar thing with a personal spaced repetition system I am building. Instead of braces I used ?question marks? and I added daily writing quota to force myself to write instead of going straight to card reviews.

https://github.com/RickCarlino/KoalaCards

(You must add cards before accessing writing)

Wow, we're on the same page. I'm making a language app specifically for couples https://couplingcafe.com ... about to release a version focused on chat that has translator support as well

How has it been since you been using your tool?

Really good idea.

I also really like the name.

Maybe add a language selector such that the AI has a better context of what language the user is interested in.

Also, yes the TTS doesn't work yet - I decided to share the MVP first before adding my Google billing account :P

My wife is Danish, we are together since most than 10 years and I never learn anything that the real basic words (tak, skol, etc.). I tried so many apps, courses, etc. to learn Danish, but always struggling with pronunciation.

I will try also that, it seems a good idea.

And good luck with your Danish learning journey :)

First of all, 10 points for the dedication and perseverance! The website looks super professional and promising. What other languages can it help with? I am planning to learn Arabic. Also, are there options to choose between different dialects of a language?

Well I broke it immediately, haha. Sorry about that. Apparently it doesn't like Chinese.

Nice one! I've been doing something similar with GPT to help learn Spanish:

Whenever I type something in Spanish you translate what I think I meant into English, point out my mistakes if any, and also respond with the corrected sentence

> because I don't know a word

You haven't used a dictionary (app)? Put the word you want in, then use one of the options in your Danish sentence and continue.

This is nice, but what I'd really like is native integration with tools such as WhatsApp. Sadly, being closed source we can't help with this.

you should make it clear that it works in any language. the name makes one think that it's a specific language (and the example as well).

though right now maybe you reached an api limit or something because any french i type comes out with an error

You just posted this two days ago? https://news.ycombinator.com/item?id=44549549

Just a small comment: The textarea has a white background and almost white text for me on iOS (dark mode), so it's not readable at all there.

I wonder with OpenAI will do with the transcripts of the chats with your girlfriend after they've been shared with them...

This is a really cool concept :) Generally in language learning it's a good idea to try to "think" in the language as much as you can so you get creative about how to phrase something with the words you do have, but if that fails, particularly when you're just starting out this is a great motivational aid and conversation starter if nothing else :) I would definitely recommend integrating with an established translator like DeepL as well as a kind of second correction as AI does get things wrong and having the two versions can help compare what may have gone wrong, you can probably keep your autodetecting language bit as well as I am pretty sure DeepL supports language identification and if not, I know Google does. Good luck with learning Danish, I tried my hand at it a few years back but have switched to Finnish as of early this year

What is the difference between this and using something like a ChatGPT project and giving it the same pre-set instructions?

Also got an error :/ but excited to lean on this as a supplement to the Hindi conversational lessons I've been taking!

Why is the suggested input misspelled? Is it because it is supposed to help with misspellings?

i tried a simple "donde (bus station)?" but i got an error about an invalid text

Just getting an error: Unable to process correction

Doesn't work for Armenian.

I am wondering if it will work for my girlfriend. She speaks English.

Super cool

Doesn't seem to work for Bahasa Indonesia, #10 most spoken language in the world[1]. What a shame!

[1] - https://en.wikipedia.org/wiki/List_of_languages_by_total_num...

Great idea and very well executed.

... Now for the hard challenge - try and speak Danish :)

wow fancy guy with a girlfriend

I’m sure she is a lovely person, but I don’t really have a need to text your girlfriend.

[dead]

I really hate the term “expat”

[dead]

[dead]