What are some good options for translating between spoken languages?
I’ve been using kagi translate https://translate.kagi.com/
Of the “free as in freedom” options: Wiktionary and patience. DeepL is not FOSS, and I suppose most other similar sites aren’t.
I’ve just looked up Wiktionary in Kiwix and it’s available to download for offline use too. Just under 8GB.
FOSS wasn’t a criterion in the OP’s request, but I’m all for FOSS options whenever they’re available. Unfortunately, for now, there aren’t any efficient and reliable tools out there :( I’ve been desperately waiting for one for the last 15 years.
This is on /c/Foss so I assume the OP was asking about Foss options for translation.
I stopped making assumptions based on logic in this world 😂 That said, I’d love to hear about any good FOSS options too. The main issue with Lemmy is that most discussions only get attention for a short time after being posted — even the OPs vanish first. :(
DeepL has served me well on desktop and mobile. Free chatgpt is not bad either.
To me, the go-to answer used to be DeepL, but more recently it’s Kagi Translate. It’s really accurate for languages I do speak (English to German and vice versa) and allows for recorded speech, although that seems to be reliant on clear talking (my first try of “Hello, what’s going on?” was recorded as “Yo, what’s going on?”)
I just tried it out and it looks great. Do you have a subscription with them?
Agree and yes! I happily pay around 100€ annually. Comes with quite a few good perks
- unlimited access to probably one of the best search engines (for search-engine-literate folk it’s not necessarily that you’ll magically find better websites, but the quantity of garbage-/AI-websites will just be heavily reduced and you won’t have the problem that the first three or so links will be ads/sponsorships; it makes searching just feel nicer)
- customisable search (for example, I can generally rank search results for pinterest lower or outright block them to not even have them show up in the first place and instead boost results from beehaw)
- customisable and privacy-respecting AI bots only if you want to use them (from what I could find out, the AI bots always start every new conversation with a blank slate of you, no matter how much you’ve used Kagi or the LLMs before)
- Kagi Small Web, which is an initiative to basically push the “alive internet theory” by highlighting small blogs and personal websites of a whole range of people
- Kagi Universal Summariser, which is an AI summariser that works very well on just about any text you give it and, for the articles and papers I threw at it, doesn’t tend to hallucinate stuff
I think the link is wrong, or at least not working for me.
Kagi does indeed look promising after a quick test. The options are particularly interesting. Thanks for sharing.
Only the requested captcha is bothering me.Thank you, it indeed was wrong. Apparently when I set the link to be translate.kagi.com it tried to send me to beehaw.com/translate.kagi.com Weird but I fixed it now.
And yeah, the options are very neat! I don’t have a captcha though, but that may be because I’m a paying subscriber to Kagi and thus logged in?
Books such as dictionaries (or dictionary websites), or we ask friends etc who speak those language
Edit: Glosbe is also a good option as it’s translated by actual people, not machine learning rubbish.
I think that URL might be for something else entirely. Unless they also do industrial manufacturing
Edit: Yup! Looks like a couple transposed letters https://glosbe.com/
Thanks! Didn’t notice the s and the b were the other way round!
I use mainly dict.cc because it’s not supporting linguee/deepl monopoly but I also use linguee if dictcc reaches its limit.
I don’t like the situation.
dict.cc looks cool - thanks for the recommendation 👍
It’s ok …
Beware it’s not foss
The problem is that it’s hard to find tools that actually handle full text well, with real context and an understanding of how language is used. It would be awesome to see more FOSS efficient and reliable tools available, but so far we’re forced to use what we have :(
We don’t even have accessible open dictionaries, afaik





