Translation corrections for Google Translate API v2

Question

Is it possible to specify translation corrections for Google Translate API v2 to use?

I'm using Translate API to translate certain Web pages, and the translation is sometimes wrong (which is expected). So I'd like to add manual translation for some pieces of text, but I don't know how to do it (or whether it's possible at all).

I know that the Google Translator plug-in (https://translate.google.com/manager/website/) allows to specify manual corrections, but that doesn't seem useful for my problem.

Answer 1

Did you complete the registration of your website through http://translate.google.com/manager/website/add ?

At the end of that process you should get a <meta> tag and some code to add to your website.

After this is done, you will be able to return to http://translate.google.com/manager/website and manage corrections and manual translations.

Answer 2

It is not possible.

As you can see here, the Python version of the Translate API is the most mature. So it stands to reason if any API will have it, that one will. They aren't going to implement a correction API until they get the "standard" translation functionality working.

Keep in mind whether you're actually using Python is irrelevant. We are just using the API as a measuring stick.

Feel free to check the Python API for yourself. It's not very big. As you can see, there is no correction capability. The only correction language available anywhere is the declarative meta tag solution already mentioned.

Speaking of that solution, the Translate team lead wrote this (albeit a year and a half ago):

"We hope this feature, along with Translator Toolkit and the Translate API, can provide a low cost way to expand your reach globally and help to break down language barriers."

To me, the phrasing of correction functionality "along with" the API implies that the two are separate.

Answer 3

You can use the Google Translate Community web page. This site let's you to send new translation words and phrases examples to Google as well as check the quality of the existing translation samples that are available in a lot of language options.

Additionally, the ML models used by Cloud Translation API are trained based on a daily usage which means every request sent to the API will help to improve the model quality and accuracy.