For members
Use Ai free on the GPU you already own.
Install the VirtualGiant client and run the whole platform on your own machines, free and without limits. When a job is too big, borrow the network with credits your machine earns while it's idle.
Credit balance
[shown in pilot]
- Your own jobs
- Free, unlimited
- Your GPUs
- 2 machines · 3 GPUs
- While idle
- Earning credits
- Temperature cap
- Within your limit
Up and running in four steps
- 01
Install
Download the signed Windows client and sign in. Add every machine you want to use; they pair to your account automatically.
- 02
Benchmark
A short test measures each GPU and your internet connection, so your machine gets work it can do well.
- 03
Set your rules
Choose when your machines lend their idle time, how hot and how hard, and which kinds of work they accept.
- 04
Use and lend
Use the platform on your own GPUs as much as you like. While they're idle, they earn the credits that pay for fallback.
Your GPU is still yours.
Joining doesn't lock your hardware to us. Run Ollama, ComfyUI, your own models, your games, anything you like. The moment you use your GPU, VirtualGiant finishes or hands off what it's doing and steps aside. Only when your machine is idle does it go back to work for the network.
You never lose credits for taking your GPU back. Stepping aside is part of the deal, not a failure.
Local first, always
Everything that fits on your machines runs there, free. Connect more machines and you get more: three machines with six GPUs means six GPUs' worth of Ai, all yours.
Small cards earn too
The network is designed so cards of every size have useful work, including on models far bigger than they could run alone. Small and mid-size cards get their share, and no single member can take it all.
Your machine, your rules
- Steps aside the moment you use your GPU, for anything
- Lends only on the schedule you set
- Temperature and power caps you choose
- A monthly data limit for capped plans
- Steps aside for updates and scans, no penalty
How credits work
No money changes hands. Credits are a record of the work your machine did for others, and they pay for the work others' machines do for you.
- Earned
- For verified work, measured by the work your GPU did
- Spent
- On fallback jobs, at the same published rates
- Quoted up front
- Before a job goes to the network you see its cost and your balance. Not enough credits? It doesn't start.
- When
- Spendable a few days after the work, once checks clear
- Your own jobs
- Always free, across all your machines
- Proof
- Every credit itemized and signed by the machine that did the work
- Cash value
- None: credits can't be sold, cashed out or transferred
What you need
At launch, every member brings at least one capable GPU. Above the minimum, your benchmark decides what work your machine gets.
| Component | Minimum | Recommended |
|---|---|---|
| GPU | NVIDIA, AMD or Intel GPU with 8 GB of memory | 12 GB or more, any brand |
| Computer | A PC that runs Windows 11 | Same |
| Memory | 16 GB | 32 GB or more |
| Storage | SSD with 100 GB free | NVMe SSD with 250 GB free |
| Internet | 100 Mbps down, 20 Mbps up | Fiber, 300 Mbps both ways, wired |
| Lending | At least 6 idle hours a day, on a schedule you set | Overnight and weekends |
Brand doesn't matter: NVIDIA, AMD and Intel cards are certified the same way, and you can mix brands across your machines. Starting requirements for the pilot; we'll adjust them as we measure. No capable GPU yet? Join the waitlist and we'll tell you when other ways in open.
What you'll never see
Other members' content. Your dashboard shows the kind of work your machine is doing, never what's in it, and identifying details are removed before work reaches you.
What we'll never do
Run unsigned software on your machine, lend it while you're gaming, or change credit rates and rules without telling you first. See exactly what the client can and can't access.
Be one of the first members.
Tell us about your hardware and we'll invite you as the pilot opens.