Macsen is an open source Welsh voice assistant app for Android and iOS devices. You can ask Macsen in Welsh language speech for information or for it to perform a task.
You can download and install the app from the relevant stores:
|
|
Macsen has 8 skills that are common in voice assistants. Some skills are not possible in commercial English language voice assistants. They are:
-
Read the news [Fideo]
-
Provide information on the weather [Fideo]
-
Play music by Welsh language artists on Spotify [Fideo]
-
Show a television program through the S4C Clic website.
-
Read the first sentences of Welsh Wikipedia articles. [Fideo]
-
Set an alarm [Fideo]
-
Tell you the time [Fideo]
-
Tell you the date [Fideo]
Along with the app's code, all of Macsen's components for speech and linguistic artificial intelligence technology are open source and permissevely licensed for use by other developers. Documentation on Macsen's software and how to adapt, expand and develop your own Macsen can be found in docs/README_en.md.
In the meantime, we are doing further research and development to improve it, and to enable Macsen in other situations.
Work on Macsen has been funded by the Welsh Government.
However, Macsen would not be possible without the work of other open source projects and communities. We’d like to thank Mozilla for its speech technologies and their excellent CommonVoice speech data crowdsourcing website and DeepSpeech speech recognition library. They simply rock.
Macsen’s intent parsing is based on MyCroft’s Padatious library while provision of open source text-to-speech is due to MaryTTS
We’d like to give a huge thanks to all the volunteers who have contributed their voices via the CommonVoice website and/or the Macsen app for enabling Welsh speech technology. We would like to thank S4C, Golwg360 (the Welsh language website), Welsh Wikipedia volunteers and the wider Welsh speaking community for their support of Macsen.
Macsen: A Voice Assistant for Speakers of a Lesser Resourced Language, Proceedings of the 1st Joint SLTU and CCURL Workshop (SLTU-CCURL 2020), pages 194-201 Language Resources and Evaluation Conference (LREC 2020), Marseille, France.