When you're using Google Translate, it sends a request for translation of your phrase and recieves a response. Of course, Translate know how to interpret response, but how about us? That's why we've made this repo. The only purpose is to show you, people, how to interpret this response and, probably, to show you a few examples of using (or even more..?)
Yeah, but how about 60-days trial and 20$ for every million of translated characters?
Price is pretty low, I must admit. What's the reason then? We're pretty stubborn, I suppose.
Developers, I guess. Response's usage is quite easy, so even a beginner with python and requests lib (like me) can use it
specification.md is what you need.
..yeeaaaah, I wasn't absolute clear with you. Google uses tokens ('tk = number1.number2' in request's data) for validating responses. So you should generate this token for sending correct request. And this token changes every hour. But the good thing about it is that server accepts old (don't know how old though) tokens.
The token is client-side calculated, so you actually can rip this code from desktop_module_main.js (relevant for 11 July, 2016) of Web Translate. But be aware, that script's obfuscated pretty hard.
It would be much better for you to check @matheuss google-translate-token for JS code of calculating this token. BEGIN-END section.
I have a lack of imagination, so if you have any questions, then create an issue. We'll add important Q&A to this readme.