Local AI Omarchy Plugin — sero.local-ai
0xSero ships a Local AI plugin on the Omarchy marketplace (sero.local-ai) and is recruiting testers to deploy local AI across more machines.
Records that name 0xSero as author. Nothing here is inferred from a screenshot.
0xSero ships a Local AI plugin on the Omarchy marketplace (sero.local-ai) and is recruiting testers to deploy local AI across more machines.
0xSero demos the Local AI Omarchy plugin (sero.local-ai) — local models as a first-class home on the desktop (plugins.omarchy.org).
0xSero previews upcoming Local AI Omarchy plugin updates: multiple GPU configs, more models, and local usage tracking.
0xSero posts a video of vibe-coding on Omarchy mid-flight using an iPhone and local models.
Long-form interview with DHH covering Omarchy as an agent OS plus Rails—widely shared video (~970+ likes) that sparked the weekend’s Omarchy conversation.
Pointer to the YouTube talk “DHH - Omarchy the Agent OS” — flagship messaging on Omarchy as the agentic desktop OS.
Demo video of Omarchy agents shaping the whole OS experience via 0xSero’s local-ai plugin (sero.local-ai on plugins.omarchy.org) — compute plugged into your preferred harness, aiming at DGX Sparks next.
0xSero asks Codex to remap Omarchy keybindings to macOS muscle memory — a few minutes later navigation matches Mac, with a desktop screenshot of the remapped setup.
0xSero demos a Battlemage B70 build designed around Omarchy as a local AI workstation — Codex, Claude Code, pi/omp, Crush, and OpenCode — inviting anyone with ~12GB Nvidia VRAM to build a box and join.
0xSero shows Omarchy on a desk at last; the photo is a wooden sit-stand desk with an Acer monitor showing a mountain-road sunset wallpaper, an RGB mechanical keyboard, two mice, a PC case under the desk, and a whiteboard labeled LLAMA behind it.
0xSero tells Warsaw friends he has two events over the next two weeks and they should join: an Omarchy meetup on October 1, 2026 (Local AI - Omarchy & Demos on Luma) and a Claude Code meetup the following Thursday; the photo is a sunny Warsaw street with city bikes and a green-domed building ahead.
@0xSero · Intel Arc Pro A70
0xSero posts security clanking on Omarchy mining bounties; the screenshot is a dark desktop with a LOCAL AI 0.0.1 overlay for Qwen3.8-27B on an Arc Pro B70 32 GB showing 44.3M tokens and 561 requests, log panes about TTSR rules and diagnose-crash skill compression, and a bottom bar reading Resume Crashed Security OMP Session.
0xSero says Local AI on Omarchy is verified now and asks for help making it better, linking the marketplace listing; the still shows an OMARCHY BAR WIDGET diagram with a Local AI hub at 127.0.0.1 keyed gateway linked to pi, Claude Code, NVIDIA., intel., AMD, Hermes, Cook, omp, Crush, OpenCode, and Codex, plus a Local AI bar-widget card by @0xSero with a green Verified badge, #782, 98 stars, and card copy that begins "Run a local model on your own GPU or CPU and open"
0xSero says "Bro deepseek is something else!": 0xSero wants to work on the design of their Omarchy local AI plugin, and it just went in there and got every screen editable with all states; the still is a DeepSeek Harness chat for omarchy-design beside a design studio page of "57 screens" set to Omarchy dark, listing screens such as home-life, home-crashed, and home-setup, with a Qwen3.8-27B card on RTX 3090 12 / 24 GB and DeepSeek V4.1 Flash selected.
0xSero says "CPU & SSD offloading now on Omarchy": "RN only supporting 3090s but I will be rapidly adding support." They think 4 RTX Pro 6000 with 256GB RAM and some NVMe can run Kimi-K3 at usable speeds, linking their plugin on the Omarchy plugin marketplace; the still is a terminal of agent output with a "TODO" list starting "I. Scaffold · 0/2" and "Qwen3.8-Flash-Next · local" in the status line, beside a panel reading "running · 36 tok/s" with "Open oh-my-pi" and "Stop" buttons, "36 tok/s" "decode avg", "10.4K tok/s" "prefill avg", and "2.2 s" "first token", a "NEEDS" list with "EXL3 3.05 bpw, vision", "sglang", "RTX 3090 · 24 GB", "75 GB" RAM, "85 GB" disk, NVMe "required", and "200K tokens" context, and "RTX 3090" at "23.4 / 24 GB" under "GPUS".
0xSero says "Local AI v6.11.2 on Omarchy is out." and "Look how I'm running GLM-5.3-Flash on a SINGLE RTX 3090 at exl3-3bpw": "What I'm doing is putting parts of the model on NVMe, parts on DDR4 RAM, and the GPU is used as a station of sorts, we send the math to be done there." They add "I value freedom"; the still is a dot-matrix drawing of two hands reaching toward each other on a black desktop above a bottom bar.
0xSero says "In August-September I started giving away hardware, I think in total I gave away 3 DGX Sparks, an AMD Strix Halo, 4x 3090s and a few thousand dollars." and "The money to do this was granted by the Human Rights Foundation, no strings attached." They add "I thought I was doing a good thing, but everything I gave came back to me 5 fold.", listing "30,000$ yearly grant from Omarchy", "Computers being sent to me left and right", and "My social media income now sustains my bare necessities and then some.", and close "I’m not allowed to lose."