Current method of storing translations is unstable and unreliable

I have been suffering from slowness, instability and unreliability in translation operations for a long time.

  1. On the portal, I rarely get the list to load completely the first time. Usually I have to refresh the browser several times until the symbol below disappears:

  2. Also, many times, when I want to save a translation directly on the portal, I get the message that it Failed to save translartion, please try again later, and I have to click on the save button a few times until I manage to save it correctly. This is a problem, because if I don’t pay attention, I might not notice that the translation was NOT saved and lose work:

  3. Once I managed to save the translation on the portal, it takes around 3 to 5 minutes for it to be active in Studio or Player. That is, every time I make a change in the translation, I have to wait this long to be able to see this result. Imagine that multiplied by dozens of translation tests…

  4. Many times a registered translation is simply not found (obviously I check the key byte by byte). In this case, I’m forced to delete the translation and reinsert it exactly as it was before for it to be recognized::

  5. Also on the portal, as reported in this other bug report, the page freezes after deleting a translation.

  6. In Studio, I notice that when I export a few hundred lines at once in a .CSV file, most of the time the translations are not entirely received by the portal. I have to repeat the process in Studio with the same file 1 or 2 additional times to get all the translations of the .CSV file correctly sent:

Something is very wrong with this architecture.

Page URL:


Hi Rogerio!

Thanks for the report! We’re currently actively investigating this issue for a solution. Could we get a copy of the CSV that you’re trying to upload so we can reproduce the issue on our end? If possible, could I also get added as a translator to your experience so I can look at the error you’re seeing?


I don’t know if it will be possible to simulate the same errors in my production environment, because as I informed, the current environment is highly sensitive and unreliable. I don’t want to lose all the work I’ve had so far because of one more failure.
Furthermore, I’ll send you in PM the .CSV and ask that if possible, I can do some tests in another area.

Edit: In addition, the sending of the .CSV refers to only 1 of the errors I reported.