current homelab setup for local AI experimentation: - hermes box hosted on a Framework Desktop Mainboard AI Max+ 395 - 5090 eGPU running Qwen 3.8 27B for fast tok/s LLM use - sometimes 150+ tok/s - 2x DGX Spark: running Deepseek v4 Flash 0731 - better but slower model - pi 5 for monitoring - Mac mini as a dev box - use Herdr and ohmypi/codex/claude depending on the use case - housed in a 10" DeskPi mini rack (mostly) Hermes is defaulted to local AI but with a homegrown routing plugin hitting a small low TTFT model (Arch-Router) to decide whether to go local or upgrade to cloud/frontier. Trying to get to 100% local over time, but right now probably more like 60-70% The Sparks are for batch processing background runs (all the cron jobs, longer dev builds, etc) do you need all of this? Absolutely not lol. I started with the mac mini and couldn't help myself but to add over time! Also I regret getting the eGPU so I wouldn't recommend that to anyone

Keep the fun parts of homelabbing. Hand off the tedious parts to an agent. Aperture’s new Tailscale and Tailscale SSH MCPs let agents add and access approved nodes without manually copying keys or dealing with configuration. See how it works →

I needed a mini mobile homelab setup for VCF Midwest, and I've never used UniFi gear before. Thus I built this tiny homelab network rack, that fits in my carry-on luggage:

This weekend I tried to let my codex rebuilt my homelab, it spent 15% my weekly tokens. 1. Fixes all known CVE in my NAS + openWRT 2. Upgrade my alist to open list 3. Implemented the inference optimization for my local qwen 4. Implemented a mobile friendly open code UI 5. Created a really nice web based book reader to read things in my NAS And then @thsottiaux resets :)

Finally got sometime to mess around with my homelab. It’s impressive on local model communities work. I can run qwen 3.8 at 110 TPS on a single 4090 now with llama.cpp. Truely goat and democratize intelligence.

i must say with as much hate surrounds @0xSero, this man simply motivates the fuck out of me by pure jealousy in me not having enough compute as him i'm going to rage upgrade my homelab so i can do more local stuff like him and reduce my reliance on frontier AI

As part of the Mega Update 🚀 Your Hermes agent team no longer has to live on one computer. You can now spread your Bots across multiple machines and bring them together inside the same Group Chat 💻🤖 → 🖥️🤖 → ☁️🤖 Your laptop. A homelab. A VPS. Hermes Cloud. Each Bot still runs on the machine where it actually lives, with its own profile, memory, tools, model, and environment. So instead of forcing one computer to do everything, you can start building an actual distributed agent team. For example: • Keep a coding Bot on your main workstation where the repo lives. • Run an always-on monitoring or automation Bot from a homelab or VPS. • Put heavier or specialized work on another machine entirely. Then bring them into one Hermes Group Chat and let them work together. Setup is pretty straightforward: 1. Add your machines under Settings → Connections. 2. Create or keep the Bots on the machines where you want them to run. 3. Open New Group Chat and select Bots across those connections. Hermes handles routing each Bot’s turn back to its own machine. And this goes a step further with gateway-hosted Group Chats. The room itself can keep running without Hermes Desktop staying open. So Desktop becomes the place you manage the team, not necessarily the computer doing all the work. Multiple machines. One Hermes agent team. 🤯

Halfway through a Claude Code project and want to switch to Codex? Then you have to re-explain how it was built, what already failed, and what’s next. ai-memory: Open-source project that turns this into a Git-managed Markdown wiki. Claude Code, Codex, Cursor, and 20+ other AI coding tools can share the same project memory — though setup differs by client. Self-host on your machine, homelab, or VPS. Use auth and HTTPS for remote access. Works for recording and search even without cloud models. 👉 #ClaudeCode# #Codex# #AICoding#