Hi Tom, there are a few options available, the most popular being Mozilla's DeepSpeech. Instructions are in our documentation: https://mycroft-ai.gitbook.io/docs/using-mycroft-ai/customizations/stt-…
The accuracy of DeepSpeech is getting better with every release but at the moment it works best if you fine-tune the model with our own training data. Once the general accuracy gets there we'll be switching over to that as the default.
Hi Matthew, there's a community contributed Docker image you can get from our [Github](https://github.com/MycroftAI/docker-mycroft#mycroft-docker-development-…) or [Dockerhub](https://hub.docker.com/r/mycroftai/docker-mycroft/)