Paulus from Home Assistant has just blogged about Rhasspy regarding the “Year of Voice” in 2023
We’ve started with intent recognition, since even with perfect audio processing and speech recognition, you still need to do something with the text! We also wanted to jump start crowd sourcing sentence templates for the IoT domain in as many languages as possible (similar to @Jarvy’s Home Intent).
In parallel, I’m laying the groundwork for the next version of Rhasspy (v3). This has been difficult, since there are almost as many use cases for Rhasspy as there are users. In general, I see Rhasspy’s goals as:
- Providing a simple, uniform API to local open source voice tools
- Enabling users to create custom voice assistants with those tools
Architecture-wise, I also need to make Rhasspy more amenable to contributions. Adding a new service is pretty difficult right now, and I think that should be one of the easiest things to do.
Here are some “what if” thoughts I’ve been having about a future version of Rhasspy. I’m curious about everyone’s thoughts
- What if Rhasspy didn’t come with a web UI, just HTTP/Websocket/etc. APIs?
- What if Rhasspy had no “plugins”, but only ever called external programs?
- What if training in Rhasspy was separated into its own standalone application?