Build a rhassy virtual assistant

Dear everyone
I’m doing an assignment as a virtual assistant Rhasspy that can answer some basic questions like what time is it or what’s the weather like today. I actually tried to figure out how to do it but the result is still nothing. I came here hoping someone could give me a detailed step-by-step guide as I’m not very good at this myself. Hope everyone can understand and give me the necessary help
Thank you very much.

Try this:

and this

1 Like

Ah, the old “get someone else to do my homework for me” ?

And if not, you could start by telling us what you have figured out for yourself, and how far you have got

No, thats bs. Rhasspy isnt that easy, at least the documentation isnt. I’ve also struggeled. Its the lack of mircophone testing abilities at the UI. It isnt just easy to spot the wrong settings.

Nobody said it is easy. That’s the nature of FOSS software - developers would rather be coding the next feature than writing user documentation. If you want user documentation and support, then you should expect to pay for it … aka commercial software.

It took me 3 starts over 12 months to get Home Assistant working, and I learned the lesson of “attempt only one learning curve at a time”. I also struggled with Rhasspy for quite a while, especially the base-satellite model … had to read pretty much all the Rhasspy documentation 3 times before it started falling into place.

Recognising that multiple “satellite” microphone/speaker units connected to a central “base” unit is logical, but it is not well explained, I started writing a tutorial … and at about page 30 I realised it’s actually a combination of 6 separate topics. It’s currently 35 pages but not finished yet :frowning: . And @synesthesiam hasn’t been seen here for a while so I’m guessing that he’s hard at work on Rhasspy v3 … which will make my document invalid. Cest la vie.

That is true Dennis :frowning: But it’s real hard to give detailed instructions without knowing what hardware and software people are using. For example, the reSpeaker RasPi HATs i use require a driver to be installed, which didn’t work with Rhasspy as a Home Assistant add-on with HA OS. Using a USB sound card or Home Assistant Supervised on Debian OS might (in hindsight) have been better options - but with their own problems.

@ThisisDennis are you currently having a problem with microphone ? Or was this just an example of something which the documentation just assumes to work out of the box ?

I don’t see how people cans seriously expect detailed assistance with incredibly vague requests.

  • What are you trying to achieve ?
  • Do you already have Home Assistant or another home automation system you want to use Rhasspy with ?
  • Are you wanting a single all-in-one voice assistant, or multiple satellites with a base station providing the CPU grunt ?
  • What hardware are you using ?
  • What have you done so far, and what is the specific problem ?

But you see, its not just “Do my homework for me”. I think the documentation could be better, that means, that someone has to write it. I didnt wrote it eighter.

No, that microphone-thing was just an example of one of my struggles at the beginning. But it could be easier if there would be just a thing like, test my speakers without tts. Just a sound. And then there could be a microphone input test as you know that the sound works correctly.
So you have 2 points to take off the list of issues.

Maybe it would be great to put in some expert-setting, cause the amount of settings in every section is way too much for beginners.
A lot of settings arent well explained, its confusing at the beginning if you have no experience in this field and dont know what you need to focus on.
But instead of discuss here, we should awnser his questions.
I hope ill find some time this days to help out a little bit, but im a little pessimistic…

I agree totally with all your points Dennis.

  • There are a lot of very steep learning curves in order to get it all working.
  • The documentation could certainly be better - not just Rhasspy, but all the other FOSS projects which it builds upon
  • Rhasspy Junior is intended to make it simpler for users to install; and now that Michael is working for Nabu Casa this will hopefully get better.

As for @Quan_Tr_n’s post…

Rhasspy is more a framework for multiple services that communicate over an MQTT broker, with goals:

  • Enable contributors to write Rhasspy apps/skills in any language that can speak MQTT and JSON
  • Allow for individual Rhasspy components to be developed independently in separate repositories
  • Make modular installations possible, with any combination of individual services

As such, it does the Speech-to-text and text-to-speech - but does not itself process the commands (Rhasspy calls them “intents”). You will need something else to answer those questions … possibly Google, possibly Home Assistant. Kind of depends on what your assignment really is.

Well Quan i am up to page 35 writing a tutorial for one possible solution … but if you don’t have the same hardware, software environment and objective then it might be worthless for you.

Have you used google to search for a tutorial or youtube video ?

If you can give us some information about your requirements, hardware and software constraints, etc. you allow us to help you