Contribute to the Voice initiative
Home Assistant Voice relies on different technologies, models, and data sources to process wake words and commands in all languages. Our and other communities’ projects rely on volunteers who donate their time, voice, or language knowledge to improve.
Note that models are trained on lots of data and that it can take time before you are able to see the impact of your contributions.
Contribute samples to Wake Word Collective
Give us a minute of your time to improve the sample database with your own samples
Donate your voice to Piper
To create a Piper voice, you don’t need an army of volunteers. You need a native speaker of a language to create a dataset, which is then trained into a Piper model and verified again by the native speaker.Check the Training guide
Are you missing a language? Email us if you want to help out! To help interested people contribute languages, we opened up the email address voice@nabucasa.com.
Other ways to contribute with samples
You can also participate in public datasets. Commonvoice and Librivox are used by many speech-to-text projects to train their models.
Become an active contributor
There are also a couple ways to actively contribute to the development of the voice initiative: