I just ordered the keyboard from Amazon as well.
Problem is, my Home Assistant server is in the basement beyond Bluetooth range. Can you think of a solution for me? Bluetooth Proxy, maybe?
The one I have is not Bluetooth its 2.4 Ghz. Luckily I just needed to run a USB extension cable to above my office door which is 10 foot from the living room.
I can’t find where to put that Yaml for the M5 atom echo.
I’m new to this and it’s obvious that I’ve got a lot of learning to do. Thank you very much for your patience with me.
Thanks. I am seeing something in people’s left menu “File Editor” and I’m guessing that this is going to be an add on I need to install.
The instructions show where to put it on installation… not once it’s been installed. I’ve tried to add the configuration to a dashboard in order to hopefully get access to the “edit” button, but it says that there is no configuration.
I’m clicking around and I just can’t find it.
I went the other way about it. I went to a dashboard and added it in. I got at the YAML, but it complained that the visual editor couldn’t handle those elements. Managed to get it saved and restarted home assistant, but it doesn’t show in Music Assistant as a speaker.
I think I need to start understanding the platform properly, rather than clicking around as I’m doing.
Still need to find a speaker though. Recommendations on general google searches are hit and miss. Need to find a compatability list.
You will not get this sorted unless you read all of what I have been linking to. You need to add esphome addon and integration.
Please read it all before asking any more questions.
I have. And I’m trying, but it says things like, " 1. click edit on the device card in the dashboard." … and the device card isn’t in the dashboard, so I’m stumped from the outset and it sends me on another side quest.
A lot of what I’m reading doesn’t work, and the videos I’m watching even only a year ago, show options that aren’t there any more, or I get asked questions that they don’t on the videos.
this line There are various methods of running ESPHome for configuring and uploading firmware to the device. Documents for these methods can be seen here for the docker command line install , here for windows, mac & Linux and here for HA add-on
in this doc esphome_firmware/M5 Stack Atom Echo at main · BigBobbas/esphome_firmware · GitHub
will lead you to this (or it should do) my mistake for not checking. It should lead to something like this Getting Started with ESPHome and Home Assistant — ESPHome.
You need to add the Esphome addon and then the esphome integration and these should help.
Chances are every answer you will ever need will be in this index somewhere.
https://jackjourneyman.github.io/homeassistantindex/index.html
Sorry again for not checking what I am telling you to read 
Thank you very much. I’ll read those.
I guess I’m occupying an awkward state… I’m trying to do DIY things, with only a base knowledge, which makes me a difficult person to help
At the moment I’m trying to find a simple target for audio which can feed a simple stereo audio out. I’m looking at the RaspiAudio devices but they seem to be hats for full pi’s which is a bit overkill, rather than a pi Zero. I do have a stereo speaker set up which is always on because multiple sources go into a simple mixer, so if I can get something that HA running on the Pi5 can see as an audio device and give a line out that I can feed into the speakers… then I’m good to go… but from my low-knowledge position, it’s not proving easy and I’m having to rely on the patience of people like your good self.
I’ve been reading others suggestions, but the talk is mostly generic about things auto turning off, etc. and I’m still on the device hunt for an actual device.
And in the midst of all this, it seems like my Atom Echo has a fault… the speaker isn’t working! So as it is newly bought I can’t take it to pieces without voiding the warantee, and the seller is telling me to contact m5stack, whose web site is blank for me, for some reason.
So I seem to have a series of problems to work against 
(I might just say “Stuff it,” and open the Echo anyway. Could be something as simple as a loose wire.)
First test your atom using this setup it is the official way to set them up. It will prove if the speaker works or not.
Do not rely on the experimental code I linked to just work.
I don’t know a lot about pi hats, but I believe the pin outs are same on all pi’s so the hats will work with a pi zero, whether it has enough power or not I can’t say. Others will no doubt know more and let you know.
What is the speaker setup you have, is there an integration for it? It may be really easy to link it up.
This is because Pi in general have AWFUL DACs and produce horrid audio.Once you add a decent DAC you basically built a hifiberry and still need good speakers.
If you’re trying to DIY for audio you will really want a device with decent output. (im resigned to buying a half dozen hifiberries for my install for music, myself.)
Kinda but. It’s OK. 
- you weer clearly warned the difficulties - I will reiterate them here. Ain’t easy at all. This is one of the hardest things you could have picked.
- you decide to continue (go you)
- we all continue helping. So we know what we’re getting into now.
I’m a very advanced user and I’m basically on the same path as you… But just to set your expectations… I started last November and have been stepping slowly along the path. I FINALLY have a working LLM driven Assist that I can. JUST Now start getting to be able to drive music. (yes just now. There are no default intents to play music, and forget searches. You have to write all that from scratch or rely on the kindness of others.) And honestly it’s a beast of a hill. (I didn’t realize I had to write the whole music handling from basically scratch.) Looking at my plan I don’t expect it to actually work until NYE. Hoping that the new voice hardware from Nabu is available. Because so far in my research (your mileage will vary etc.)
Box 3 hardware is great to test wake word but there some serious limitations most notably around the mic array and the ability to continue conversations in the chat context. And the speakers onboard are simply awful. You Can redirect the voice response pipeline to better audio devices but now you also have to handle that.
I am in no way trying to discourage you but I want you to know the difficultity level and the number of issues you have to solve to make this work.
Choose hardware (prob start with a Box3 to learn the ropes)
Buy and install hardware
-includes learning ESP32
Install and build assist.
Install music provider
Get assist working to play music, skip next. Previous. Etc. (all of these voice commands meet to be built)
Install search provider for your music if its not provided (hint it’s not…)
Write handlers to perform search for artist. Genre, etc.
Get assist working with all those.
Get the llm working with he assist controls (yes you’ll probably have to explain what the llm sees to it)
Now you can TECHNICALLY ask for playback - and see how much you now need dedicated music hardware.
Thats your hill to climb. I’m somewhere in the get assist working to play music with new intents phase.
That said, when you get it working you also have completed a masterclass in Home Assistant - -proceed tackling any challenge.
(and i would 100% never ever use this with my mom. In it’s current state. For her she gets non ai Alexa and a Sonos.)
1 Like
The Atom came pre-installed with the right firmware for Home Assistant. That seems to be how they are shipping them now. I’ll keep that link up my sleeve and see what they come back with.
I’ve sent raspiaudio an e-mail because they’ve got a product that would be great for Mum… the radio. She’d have a nice enclosed system rather than the raw boards I put up with
The audio system is a simply a resistive mixer (USB powered, but non-amplified) that goes off to a pair of Logitec speakers. It takes my BBC Micro, Original Xbox, MiSTer, etc. into a single output. Being dumb speakers, they don’t turn off 
Thanks Nathan,
Yup… I do things the hard way usually
I have added the voice control script over to the Pi, but I’ve got no working speaker so I’m a bit hosed for the moment - Voice Control - Music Assistant
I’ve also now got two ReSpeaker Lite units programmed. One is two meters away from me… behind me… and still picks up my speech. Stunning.
I was able to program those because I did a wi-fi storage project on my ZX81 computers using an ESP-WROOM32 so it turned out that I had the esptool necessary to blow the images to them. However, in the mean time I accidentally blew the firmware onto my Sonoff dongle, and I can only get the repeater software to get back on to it in a working state… so I’ve got to wait for another one to come before my Zigbee2MQTT is back up and running.
So once I’ve got a viable speaker that Music Assist can see, then I can test the code that I got and see how well voice control works… and then… I can move forward to the next step.
I used to program back in the days of 6502 assembler, BASIC, Pascal, COBOL, and lately I learned PHP… but that was years ago and I never did anything that was really object orientated. The trick with me, is getting the penny to drop. Once that happens, I’m off and running… but I’m under no illusions… programming the ESP chips is nothing like the 8 bit machines I grew up with, and the Home Assistant also has its own way of working that I need to get to grips with.
1 Like
K the in that case your hard hill is the voice intent.
The hardware will frankly resolve itself. “it’s just time and money” applies.
Spend your time studying how the voice pipeline works in assist. Understand how intent and Intent_script works together to make things work. This has by far been the hardest part. (but I swim in small electronics and picked up esp32 stuff quickly)
When testing intents I’ve found it’s helpful to have two voice pipelines setup nearly identical except one is driven by assist and the other by my LLM of choice. Get your function working with assist first Then add the llm. Watch it fail a few times and then figure out how to edit the prompt. helps figure out where your issues are in debugging.
Good luck and I interested in what yih ultimately choose on your HW.
1 Like
Well… from what I’ve seen so far, I don’t think that the ESP chips I have will handle a stereo audio stream with any grace, so it’ll probably be Pi Zero.
A Pi Zero is the basis for the RGB to HDMI project that is used on the BBC Micro to convert its digital RGB signal to HDMI (with a custom board) so I do believe that coupled with a RaspiAudio sound card (the small one) I should be good to go for something that will take from Music Assist and feed out to a stereo jack. That’s number 1 down… I hope. I’ve got one coming, but I think it’s from the US, so it will have to cross the pond. Delivery will be about three weeks according to Amazon.
That will give me time to find another device to test with… and if I don’t get any joy early next week on the Atom then I’ll just take it apart and troubleshoot… but I’m still on the look out for another low cost audio player that HA will just see and use, so I can test with in the mean time.
The mics on the ReSpeaker Lite, I’m impressed with. I have put one in a home made shield and that will stay in the office. It’s hooked in to the Logitech speakers for announcements… shame I can’t easily use it as a target for music. At least… that’s the way it looks.
I’m a bit cautious about what scripts I’m putting where, as I dread toasting the thing, so I’m only really following instructions that tell me where to click… because I know enough to know that I don’t know enough
It’s Bluetooth (BLE).
But, Home Assistant doesn’t discover it through the BLE Proxy, so the keypad is of limited use to me.
Sorry about the off-topic fork. I’ll end it here.
Where can I get the mixer?
@stevemann I really don’t wish to derail this topic any further, but catch this post of mine, in case it helps