Call for cooperation in making multilingual sentence decks

I’ve been working on the following project

  1. I’ve gathered all the shared Anki sentence decks and merged them into one file and cleaned up all the cards.
  2. This has been done for the following languages (Portuguese, Spanish, French, German, Russian, Chinese). They all usually have English translations.
  3. Using deepl translation services (which gives excellent machine translations) I’ve added (or plan to add) translate all the above mentioned languages into the others.
    Note: Other common languages such as Arabic or widely spoken languages such as Malay/Indonesian or Hindi/Urdu are not added yet because deepl does not offer translation in these languages and other machine translation services are not as accurate.
  4. In the tags one can see what the original/base language is (and thus the sentence should be free of grammar mistakes).
  5. Now through AwesomeTTS add on. I’ve added azure neural text to speech sound for each sentence.
  6. Finally, through an algorithm I’m created I will add all the notes together and sort them from easiest to most difficult sentence based on word frequency and sentence length.

So far I’ve made good progress for Portuguese which I’m learning right now
a) Out of the 161,000 unique sentences over 130,000 have had speech added.
b) All sentences have been translated into the aforementioned languages.
c) Only a couple of thousand have audio from all the aforementioned languages.
d) English sentence sounds need to be added.

Here is the problem I’ve encountered

  1. Aside from being very time consuming for what will be a corpus of about 1 million sentences it’s also quite expensive using deepl, and AwesomeTTS. Currently, a premium subscription of $20 will only allow one to translate about 20,000 sentences a month. As such, just adding audio to one language could take a year.
  2. Deepl also has limitations in the number of translations it can do though it should be possible to translate a deck like the portuguese in a month or two. When translating large volumes of information from a spreadsheet there can be glitches and you will have to scroll down the spreadsheet to check the translation matches.

So my idea is that this project could be shared out amongst a number of people. We could cooperate and create one of the finest language tools on the web. Who’s will me?