The Wikimedia Foundation is currently developing a speech synthesis engine to its encyclopedia. In parallel with the release of its new app for iOS , the Wikimedia Foundation is currently developing a multilingual engine in partnership with the Royal Institute of Technology KHT. It will be optimized for Wikipedia but available as open source.

Joakim Gustafson, professor of speech synthesis technology said in a statement that the project aims to facilitate access content from Wikipedia to the visually impaired or have difficulty reading.

Mr. Gustafson added that the first tests will be carried out in Sweden. He added, “then we will do a basic English voice, which should be pretty good since we have full linguistic resources in open source.” Eventually the work will be carried on a voice in Arabic.

Despite this, people with blindness and illiteracy are often excluded from online news, information, and educational materials. But with a text-to-speech API provided by companies such as WellSaid Labs (https://wellsaidlabs.com/), and other such companies in the speech synthesis sector, it might be possible to quickly integrate audio-based content, which can be consumed by a much larger population, making it possible for people to access the same information as everyone else.

Coming back, the Wikimedia Foundation estimates that 25% of users – 125 million people per month – prefer or need a voice device for reading texts. These first tests should be finalized in September 2017 and will be followed by other work for other languages. The service will be offered.

Wikipedia plans to have English, Swedish and Arabic ready by the end of the year before moving on to the other 280 languages its site supports.