Mini PCs, GPUs, storage, and Home Assistant sensors — the exact gear behind the guides on this site. No sponsored fluff, just what works.
The sweet spot for running models at home without a cloud bill.
Fanless, ~$150, sips power. The perfect dedicated Ollama box for 7B-and-under models.
Check price → ComputeSlightly faster than the N100, still silent and cheap. A good step up if you want a little more headroom.
Check price → Power pickiGPU via ROCm, 32GB DDR5. Runs 7B models beautifully without a discrete GPU.
Check price → ComputeThe swiss-army knife of the homelab — DNS, Pi-hole, light services, and tinkering.
Check price →Where real model throughput comes from — and the RAM that feeds it.
The best dollar-per-token GPU for home LLMs. Used market sits around $200.
Check price → Future-proof16GB VRAM opens the door to 30B+ models. The upgrade path when 12GB isn't enough.
Check price → EssentialThe must-have upgrade to escape the 16GB bottleneck on mini-PC builds.
Check price →Models load from disk first — fast NVMe cuts cold-start time dramatically.
The gear behind the Home Assistant automation guides.
The official Zigbee/Thread coordinator. Plug it in and your whole Zigbee mesh just works.
Check price → ZigbeeA proven, cheap Zigbee radio with a big community. Great pairing with Zigbee2MQTT.
Check price → AccessoryMove your Zigbee dongle away from USB3 ports to kill interference. Small fix, big stability win.
Check price → SensorTiny, reliable, and cheap. The backbone of door and window automations.
Check price → SensorAccurate climate data for thermostat, humidifier, and aquarium automations.
Check price → SensorFast, low-latency motion detection for lighting and presence automations.
Check price →