The APIconomy for the Language Industry
Have you seen that IBM has opened its language, vision, speech and data APIs to developers so that they can build innovative products and services on top of them? Since its incorporation, Google (now Alphabet) has invested in APIs, see a sample here, and even Microsoft has also an API catalog! Application Programming Interfaces (APIs) are everywhere.
Where are we in the language industry? When we started several years ago, the pioneers in the field where translated.net and Worldlingo. Since we were developing mobile machine translation solutions, they offered the easiest way at that time to plug machine translation solutions into any other system, such as our image translator. And from what I see, they are still among the top innovators, together with some other companies such as lingo24 and gengo.
For tauyou language technology, 2015 was the year when we started and launched our API. At the beginning of the year, we created a post-edited machine translation API to allow any company in the world to request not only machine translation services, but also human translation through our platform, powered by a community of more than 1500 freelance translators that signed up. It was an interesting learning process, that allowed us to have one of the best years at tauyou. If you use an API and try to reach companies with internal technical teams, integrating translation services becomes just a matter of one line of code. You change then from selling words to solutions, and can play with innovative business models to satisfy the needs of your customers.
After that, we wanted to make the use of our services as easy as possible. Therefore, we used our APIs to build the integration with major CAT tools: XTM, Memsource, SDL Studio 2011, 2014, and 2015, memoQ, Wordbee, and Matecat. Also, we have integrated our services with proprietary TMS systems and CAT tools from some major translation companies. The process with the developers was always easy and well managed, and whenever an issue was found, the reaction of the companies was extremely fast. From this process, and their integrations with TMS tools such as XTRF or Plunet, I feel that the language industry is prepared to reduce the fragmentation and allow companies to use any technology and service within just one platform.
Regarding the integration within other systems such as CAT tools and TMS, the advantages for our clients are clear: they use our machine translation services through their preferred tools without going to a third party environment, and we handle everything below so that engines are updated periodically thanks to our automatic post-editing rules. We tried to keep it simple. And data speaks for itself: the usage of our machine translation services exploded in 2015!
In 2016, we want to open our systems much more. Besides some pending integrations such as Across or SDL WorldServer, the objective of the year is to allow any developer to use some of the techniques we apply within our machine translation engines, e.g. recognition of places, people, companies, and events, Part of Speech tagging, tokenization, sentence reordering, synonym/antonym proposals, topic classification, sentiment analysis, automatic post-editing, summarization, and many others. In this sense, if you have an innovative service you would like us to integrate and open to the world, just let me know.
If the translation world is becoming every day more technological, any LSP should be thinking this year about the API strategy to offer their great linguistic services to end clients. What are you waiting for? If you don't know how to start, my session at the GALA Annual Conference will try to outline a step by step guide. Will I see you in New York?