Omarchy 4.6 is not just leaning into local AI, but will let you share any model running on any Omarchy machine on your tailnet with another. Get a DGX Spark, then run inference from your XPS13 netbook. Sharing is caring.
Image from DHH (@dhh)

Nov 11, 2026 · 3:14 PM UTC

182
132
35
2,847
76,997
RelevantRecentLikes
My poor local clankers have answered this test prompt so many times they must be getting sick of riffing on Pax Romana!
Image from DHH (@dhh)
9
121
14,443
Sweet.. I have been tinkering around the edge of this with gliff / tumx / filesystem on a NAS but not got it perfected. I am sure you will improve on my cludges. We need a headless Omarchy log in also.
1
34
Ok listen I posted this public statement because I disagreed with your approach but DAYUM that is SEXY UI you got there 😍😍
1
41
Mr Hansson, when will we get omarchy server? I have been running omarchy on my pc for over 420 days now, and I just installed it on my @dell latitude. I love it. However, I really dislike my Ubuntu server (esp. after the TCP RCE scandal). When can my server become great?
Image from Tsurg ↗️ (@tsurg_)
Image from Tsurg ↗️ (@tsurg_)
9
813
Can we also get a NetBird ?
10
this is how families should run ai, one box at home with the model and everyone's data, every phone and laptop just talks to it
25
1,238
Awesome!!! It will replace my local hack
20
the omarchy mesh would be awesome to “share” inference across the community
2
18
906
love it but I don't have the resources to get a spark, but I am having fun making it work on potatoes lol
31
Man, I love where Omarchy is heading. Computer use capabilities are the thing I’m waiting the most though. I just want to talk to my phone microphone with grok voice and get things done.
1
1
300
Any good local model suitable for a Beelink SER9 w/ 64 GB RAM?
42
About that. Can you send me one those sparks when you're done with it?
87
It's a great deal no doubt, but you could have added any form of installing into existing partitions, like an ext4 home or fat efi partitions!
99
Best method for having multiple agents coordinate in Omarchy? Is it Herdr? Does Omarchy based agents work better when a specific harness?
1
2
596
Love this. No other OS is this cutting edge.
135
But can it run doom
254
@dhh is tailscale coming preinstalled by default now?
104
now i just need money to get a few DGX sparks lol. I am currently on the claude max plan and im always trying to find ways to be efficient with my tokens
64
Omarchy running on any potato I throw at it has me rethinking whether I need a high-end laptop at all. In the end, a good screen, enough memory and long battery life is all you need. Everything else happens in the homelab.
1
122
That would be something else, still can’t calm down from the 4.5 hype and now 4.6 got me sweating!
1
334
Love the feature, but please fix the bugs, tiled and docked managers are still buggy.
173
Security? We don’t need no security! — DHH
56
I'm all set up except money to buy DGX Spark. Do you mind sharing yours? 😅
88
Just my thinking. Next would be some kind of advice on which model I should/can run on any new hw added. Like “Omarchy, suggest which model/LLM that shall run on my local hw - matching the projects/tasks/chats that I do”.
1
5
1,086
can two machines use the same model at once?
137
sharing models over tailnet is decentralized inference vibes
94
I do think Omarchy will have my attention for a long time more than any other Linux distro ever could. If it ever replaces or looks at replacing the guts like systemd, gets things like AppImage I will have any even harder time looking away.
5
1,076
what api do they need to serve?
264
Great to see
6
864
this is what makes local AI practical for most people. one powerful box does the inference and every other machine on the tailnet just uses it. you buy the GPU once instead of putting one in every laptop.
6
764
Awesome addition to already great Omarchy OS. You are doing great work.
274
sure, let me borrow the 8k for a spark...My money tree died
183
does the netbook get any say when the spark is already busy loading a different model? that's the case i'd want to see, two people asking for two models at once
219
Will I be able to use the inference on my Mac Studio? If not… I guess I’ll have to build it.
184
Nice, I'm currently doing this with LM Link.
1
350
When the netbook gets an answer back does it know which model on which machine ran it?
3
677
Load more replies