Provider hardware guide · NVIDIA DGX Spark (128 GB unified memory)
A DGX Spark has 128 GB of memory shared by its graphics chip and CPU, enough for large open models on one small box. We are opening model hosting on Sparks: join the list now and yours is reviewed and invited as hosting for it opens.
| Workload / Benefit | How it runs on DGX Spark |
|---|---|
| Model hosting, after review | Large open models fit in 128 GB on one box. Hosting for Sparks opens after a short review; join /host-models to be invited in order. |
| Not media jobs yet | Our media workers are built for regular NVIDIA PCs today. A Spark takes no jobs until hosting opens for it, and we will tell you before it does. |
| Runs beside your own work | You choose when the node is online. Customers never get shell access to your Spark. |
Sign up, then join /host-models and pick DGX Spark so the review knows what you have. Nothing runs on your Spark until hosting opens for it.
Pick your card's video memory and demo-hosting preference to see eligible work types and your tier from the live payout ladder. No hourly rate is promised—every finished job shows on your dashboard.
Bonus: When your node is online, your own legit2clip jobs run free on your own GPU first.
Free to use. Paste this into any page, wiki or blog post. It shows the live payout ladder every time it loads, so there is no table to keep up to date.
We do not promise a fixed hourly rate because payouts depend on real customer demand and completed jobs—never fake hash-rate numbers. Every finished paid job is checked and credited on our live payout ladder (from 5% to 25%), and you can boost your share by hosting optional free user demos.
Not with away mode enabled (the default). legit2node stops accepting new jobs the moment you move your mouse, type, or launch a game, and runs strictly inside Docker containers.
Verified earnings clear after a 14-day hold. Redeem them as platform credits with a +25% bonus or withdraw via PayPal, bank transfer or crypto (10% fee).
Yes. That is what a Spark is for here: join the list at /host-models and it is reviewed and invited as hosting for Sparks opens. Until then it takes no jobs.
Other NVIDIA GPU guides: RTX 3060 12GB · RTX 3090 24GB · RTX 4070 / Ti · RTX 4080 16GB · RTX 4090 24GB · RTX 5090 32GB · 48 GB RTX A6000 / 6000 Ada · RTX PRO 6000 96GB · Salad alternative · Vast.ai alternative · All GPUs.
Free to join, paused whenever you're using it, and a bigger share of every paid job when it hosts free demos.