Macsen
Macsen is an open source Welsh personal assistant software similar to Alexa or the Google Assistant.
Open source means that anyone can see, modify and distribute the code as they wish. It works as an app on a phone or tablet, and is available for iOS and Android devices as well as an online version. It is possible to speak to Macsen in natural Welsh in order to ask him to complete tasks or ask for information.
Macsen’s Skills
By now, Macsen has a number of skills, including the ability to play Welsh music on Spotify, programs on S4C Clic, turn the light on or off, recite the latest news from Golwg360 headlines and providing weather forecasts from a website from the OpenWeatherMap website. Macsen also uses ChatGPT, translating English speech into Welsh text and transcribing Welsh speech into text. So it is possible to type as well as use the voice to ask questions and give orders to Macsen.
Assistant
The latest developments in the field of artificial intelligence have changed the face of the field, and caused us to reconsider what it is possible for a computer to achieve. By now, Macsen uses the ChatGPT-4 Language Model to answer questions, reason and chat through the medium of Welsh.
Translation
Macsen’s translation skill enables users to translate English speech into Welsh text. By now, the technology is also linked to the Education Termiadur dictionary, which provides a useful feature in order to list terms related to what was spoken. Our other resources available to facilitate translation such as the aligner and our resource for sharing translation memories can be found here.
Transcription
The Macsen transcription skill enables users to transcribe any Welsh speech into text. After speaking your message, you can copy the text into any app, whether it’s a text message, an email or a shopping list! As well as a skill in Macsen, the software program Trawsgrifiwr is also available as an online version and as a Windows version.
Download the Macsen app
Be part of the Welsh language’s digital revolution and download the Macsen app for iOS or Android devices today!
Welsh language technologies within Macsen
Macsen uses a number of different technologies in order to work. It uses Mozilla DeepSpeech speech recognition to translate what you say into text. Following that, intent recognition technology is used to identify whether that was a request for news, the weather, music or one of the other options. When Macsen needs to respond verbally, it does so by using text-to-speech technology to speak the appropriate response.
We are still improving the speech features, and if you would like, you can help us improve it in the future by contributing recordings of your voice. You can do this within the app by clicking on Training. This will guide you to read sentences identified for the skills in the app aloud. We will use these recordings to create development sets and test sets for training the speech recognition. If you want to contribute more than this, visit the CommonVoice Mozilla website to record sentences for the large collection of recordings. More information about these technologies and the Welsh language is available in the Handbook of Language Technologies published by the Coleg Cymraeg Cenedlaethol.
Macsen and the speech recognition work were funded by the Welsh Government, and we thank them and the volunteers who have been contributing their voices to improve speech technology.
Thanks also to Golwg360 and OpenWeatherMap for permission to use their online services.
Open source resources for developers
We are using this project to show what we can create when developing speech technology and Welsh artificial intelligence. All the relevant components and resources are shared below under an open source licence:
Requires: Code to the app / Code to the parser / Documentation link