Homelab Twin · 3D rack planner
Drop in a spreadsheet of your gear. See what every device draws and costs, what's plugged where, and what goes dark when one thing fails.
Try it free — coming when we open the doors; get notified. Your file is read in your browser and never uploaded.
Recorded from the app. Every control here swaps a clip.
Lights-out closet
−145W
−€34 / month
Room light off, rack still running. Homelab Twin counts the light as a load like any other and shows the saving next to the watts. estimated
What it shows
Watts per device and money per month at your electricity price, estimated from each model and any readings you add. Follow the power from the wall through the UPS and power strip to every device, with the load on each step.
wall → UPS → power strip → device · load %Router to switch to device, port by port, in 3D or as a clean diagram. On Pro, trace every cable at the back of the rack from end to end, through the patch panel.
ISP → router → switch → device · speed · PoEClick the UPS, a power strip or a switch and press Fail. Homelab Twin shows what loses power or network, and what keeps running. In every plan.
Fail → what goes darkHow it works
Or start from the Basement Lab example.
Rack, U and model are the only required columns. Add cables, power and readings when you want more.
You get an import report and your rack in 3D. Unknown models still show up, as plain boxes of the right height.
.xlsx, or one CSV per sheet. The file never leaves your machine.
Views
The whole lab at a glance: racks, devices, and the gear outside the rack. One readout on top: watts, cost per month, UPS runtime.
Overview, Power, Heat, Network, Cables, Compute and AI sit in one View menu (keys 1–7). Power and Network switch between 3D and a diagram.
Single points of failure, overloaded strips, UPS runtime at peak and more, ranked with the fix. Try runs it once and opens the first issue.
Every cable at the rear of the rack, coloured by type and traced end to end through patch panels.
Tokens per second for the AI box in your rack; tokens per kWh and cost per million tokens on Pro.
Checks
Every import ends with a report: rows imported, rows skipped (with the reason), and checks pinned on the rack.
Share
Every lab gets a ready-made share card (1200×630), a clean rack-elevation PNG, and an animated share: a short GIF for forums and a video of your rack with the LEDs blinking and the power and network flowing.
AI
Tokens per second tells you how fast it is. Tokens per kWh tells you what it costs. Homelab Twin puts both on the rack, next to the watts.
Work out your tokens per watt with the free calculator →
The AI view's tokens per kWh is part of Pro. The calculator is free.
Pricing
Learn
Built by data-center people. Homelab Twin comes from a team that builds 3D models of full data centers, scaled down to the rack in your basement. In partnership with Sensaka.
Try will be free, with no account and nothing stored. It isn't public yet: for now we show the app in clips, stills and walkthroughs. Account ($5/month or $50/year) keeps your lab; Pro ($20/month or $200/year) adds cable tracing, tokens per kWh, the full Flaw Finder and reports. Account and Pro are launching soon.
Not on your own yet. Watch the clips on the See it page, or request a walkthrough and we show you the app on a short call. Meanwhile the import template is free to download and fill in.
No. It runs in your browser and imports a spreadsheet: the template, or one CSV per sheet. Only the Devices sheet is required (rack, U and model).
No. The spreadsheet is read in your browser and never leaves your machine. Without an account nothing is stored: no cookies, no browser storage, and no requests to third parties.
They are estimates: each device's typical draw from its model, adjusted by any readings you add to the Readings sheet, times your electricity price. For exact numbers, add readings from a smart plug or your UPS.
One or two racks, including 10-inch mini racks. Gear outside the rack, such as access points, cameras or the ISP box, can be placed too: leave the U column empty.
How many AI tokens your hardware produces per kilowatt-hour of electricity. It is the efficiency number for a home AI box, the way miles per gallon is for a car.