We’re staying compatible with existing community built skills.
Running Neon AI’s OS, “shut down” will power off the fan and screen, and leave the device in a lower power state where it is safe to unplug or remove the flash drive.
Input welcome on dev targets -
Neon AI is committed to Open Source, a long time Mycroft Channel Partner & OVOS contributor. I’m excited to unbox my own Mark II tomorrow. Getting hardware produced and shipped is a huge accomplishment, especially right now. Thank you Mycroft team!
Awesome - congrats on the release!
Excited to see more images available and provide more choice to the community
I have gotten this installed on my Mark II. It boots, and very simple stuff (like time and weather) works. Is there a place where this new version is documented or where we can ask questions specific to Neon? I just need some basic stuff first, such as how to log in via SSH and how to shut down. This version appears to run ssh on standard port 22 (which I discovered with an NMAP scan), but I don’t know the username and password to use (I tried “mycroft”, “neon”, and “root”, and all ask for a password; I have installed my RSA key in MyMycroft).
I’m thinking maybe I need to just delete the old device and start all over. Is there any documentation on how to get started with Neon on the Mark II?
I did finally figure out with trial and error that in order to activate Neon, you have to say “Hey Neon” instead of “Hey Mycroft”. I suppose they figured this was obvious but it wasn’t to me.
Hi Greg, I’m glad you got it figured out, I’m working on more explicit documentation right now.
Yes, we have a more specific area about Neon AI - I’ve set up a public “room” on Matrix, and messages posted there will be visible to our whole team, giving you the best chance of a swift answer from whichever of us knows that area best. Here is a link -
Our community: https://matrix.to/#/#neon-ai-community:matrix.org
Neon AI for Mark II room: https://matrix.to/#/#NeonMycroft:matrix.org
If you’re not up for corresponding on Matrix, you can normally get a quick answer from us on here as well. We all took a break for the holiday weekend, but are back on deck now. You are also welcome to email me directly, email@example.com or our general support email: firstname.lastname@example.org
Some suggested commands for Neon AI should be displaying on the screen, and there is a skills list on our main website and a good bit of documentation on Github. We realize that’s not ideal for end users and I’m working on much nicer “how-to” documentation.
For you ssh question, I need a more technical person - @djmcknight358 or @JarbasAl can one of you help?
I really appreciate your input, an outside perspective is always valuable. I will make sure that the wakeword / attention getting phrase is explicitly explained in the documentation, and welcome any other suggestions.
Thanks for your feedback here; all good questions that we should have answers for on the website.
For SSH, we do run on the default port 22 with SSH enabled by default. You can login with user/password
neon and should be prompted on first login to change the password. By default, we don’t pair with the Mycroft backend, so the old device listed there is probably linked to the first activation with Mycroft/Dinkum installed.
Hi, so I was able to connect and I would like to view the coding of a skill. What folder are the skills in, does each skill have a separate folder like in the prior Mycroft builds?
Great! Yes, separate folders for each skill. Here’s our Github, and I’m tagging @djmcknight358 in case he wants to add anything.
I’ve updated the Neon Docs to include some more information about where skills might be installed. Mycroft “classic” skills are fully compatible and the default path to install those to is
~/.local/share/neon/skills (XDG location as proposed in mycroft-core #2803).
The link in the original post is giving a 404 error.
Is there another place to download the NeonAI image? Is there a video anywhere that shows what NeonAI looks like and maybe what works for MyCroft?
Edit: Purchased the USB key. Some really neat videos here. Hear Conversation Processing Intelligence in Action | Neon AI
Thank you for taking the time to reply @blarghom!
- Here is the current link - Neon AI OS for the Mark II | Neon AI
- I’ve set up a redirect so the old one will work also now.
- I’m crawling the site again for any other broken links.
- Glad to hear you liked the videos! We’re working on more. Videos seem to be even harder to keep up to date than documentation.
- I’ll send your USB right out! Pleased you found that helpful.
Ok… so Neon AI is compatible with Mycroft Mark I skills…
@jmillerv has set up an initiative to convert Mark 1 skills to Dinkum here
I see Neon AI has figured out how to shut down the Mark II gracefully, as well as how to produce a list of installed skills, even if only spoken.
Am I saying something unFOSSlike when I suggest these particular skills should be forked from NeonAI onto the Dinkum repository, so time can be saved?
addendum: I do intend to order the Neon AI USB, as I did get two Mark IIs, and why not… I’m not trying to keep everything Mycroft, but just see some possible timesaving options… Maybe because programming still eludes me
I have actually have a mark II dev kit that I’m using with NeonAI. As soon as I print the Mark II case, I’ll functionally have two Mark IIs.
I’ll make it a goal today to start poking inside of NeonAI. I assembled the dev kit last weekend.
Hi All, I installed Neon on my MarkII, but one thing that confuses me is that the osm command is very very very slow (takes like 1 minute 30 seconds to respond to just “osm” with no arguments). Anyone has an Idea why that could be ? I realize that neon-skills.py and ovos-shell seem to pull lots of cpu… but is that all fine ?
I just tried to install something on my MKII
Empty command takes just over 1 minute
Almost looks like it has some kind of time-out before finishing the command.
Where do I find documentation about this.
Google got me here instead to any official documentation.
We have documentation here. Those should be fairly complete for the Mark2 and there is more documentation on the way .