Lingoes's word translation function is very likely to leak privacy

I have always suspected that the translation content requested by the user is transmitted in plain text. Today I installed a local proxy Privoxy to confirm this speculation.

First, point the lingoes proxy server to Privoxy:

Then enable the lingoes word translation function (translate selected text ):

 

When you are editing a piece of hidden text that you accidentally select, lingoes will try to automatically translate the text:

From the logs of Privoxy, we can see that it is transmitted to several translation servers in plain text:

Privoxy logs

Jun 04 18:23:03 Privoxy (00000604) Request: api.dict.cn/wapi.php? Q = FTP % 20 server % 20% 3A % 20myserver.com % 0d % 0 aaccount % 3A % 20 administrator % 0d % 0 apassword % 3A % 20 p % 40ssw0rd123 & client = lingoes
Jun 04 18:23:03 Privoxy (000004f4) Request: dict.hudong.com/dict.do? Title = FTP % 20 server % 20% 3A % 20myserver.com % 0d % 0 aaccount % 3A % 20 administrator % 0d % 0 apassword % 3A % 20 p % 40ssw0rd123 & from = lingoes & type = 1
Jun 04 18:23:03 Privoxy (2017177c) Request: www.google.com/uds/gtranslate? Callback = Google. language. callbacks. id100 & context = 22 & Q = FTP % 20 server % 20% 3A % 20myserver.com % 0d % 0 aaccount % 3A % 20 administrator % 0d % 0 apassword % 3A % 20 p % 40ssw0rd123 & langpair = % 7czh-cn & Key = notsupplied & V = 1.0

Who can ensure that the owners of these servers do not collect, analyze, or even illegally use the customer request data?

Who can ensure that the plain text transmission will not be intercepted by sniffer on the network?

Therefore, please use lingoes's word-based translation function with caution.

Here, I only analyzed the lingoes word translation function, because I think it is the most serious security problem. In fact, even if you specify the translation with the mouse, there are also different degrees of danger. At the same time, I believe that other online dictionaries are very likely to have similar risks.