However in the same time I can see that some active development is happening by Michael Hansen in other repositories synesthesiam (Michael Hansen) · GitHub
I know Michael has moved to Home Assistant, so his main dev focus will be around that project.
I wonder if there’re any forks of rhasspy been made and community has moved away to some single project? Or just dispersed?
I’m happy to put any volunteers on as maintainers for the various Rhasspy repos. I archived them because I was heading towards burn-out if I didn’t take something off my mental load
At this point, the Wyoming protocol and the voice subset of the ESPHome protocol are doing a lot of what Rhasspy did. I’m doing my best to ensure that the various Wyoming services are available as regular Docker images outside of Home Assistant so that Docker compose can be used to recreate what the Rhasspy web UI used to do when you’d pick the various services. Most of the services automatically download their models on first run, so you can get something running pretty quickly though it is less intuitive.
One piece that is definitely still missing is the server “brain” itself. The voice pipelines are run in Home Assistant, and there isn’t a Wyoming version of the pipeline runner that exists as a standalone service (with satellites connecting to it). With Claude, I think this would be a pretty quick project, but I haven’t tried it out yet.
Are people mainly interested in using the voice services outside of Home Assistant? Or is it the web UI that is missed more?
