Sign in, add prepaid credit, and set up your SSH key. Choose a DGX Spark when you are ready to deploy; credit does not reserve capacity.
Connecting to GPUwerk…
Checking capacity…
Pricing & billing · Setup guides · Service status
Need help choosing? Discuss your workload.
This is the offline demo of console.gpuwerk.com. Everything below runs from this one file, state lives in your browser localStorage, and nothing here reaches a real server.
EU-Central · checking availability · first top-up: pay $10, get $20 in credit
Every instance is one whole machine. Nothing is shared, virtualized, or oversubscribed.
Unified memory is the whole point, a single node holds models that would otherwise need several discrete GPUs.
Per-minute billing, $0 setup fee, no egress charges. Stopping reserves your Spark at a discount off the hourly rate, terminating is what stops billing. First top-up: pay $10, get $20 in credit — and refer a friend for another $50.
Every instance is a clean box with a hostname and key-based SSH. Root on real hardware, bring your own stack. Managed model serving is coming soon.
# deploy from the API curl -X POST https://api.gpuwerk.com/v1/instances \ -H "Authorization: Bearer $GPUWERK_TOKEN" \ -d '{"image":"dgx-os","ssh_public_key":"ssh-ed25519 AAAA..."}' # your shell on the box ssh root@my-instance.gpuwerk.com # the whole GB10 is yours nvidia-smi && python3 --version # locked yourself out? reinstall every key on the account, no rebuild needed curl -X POST https://api.gpuwerk.com/v1/instances/<id>/reset-access \ -H "Authorization: Bearer $GPUWERK_TOKEN"
The short version of how billing and instances work.
Per-minute, against a prepaid credit balance. Stopping an instance keeps its Spark reserved for you at a discount off the hourly rate; terminating releases the hardware and is what actually stops billing. Your first top-up of at least $10 earns a matching bonus of up to $10 (pay $10, get $20 in credit), and if your balance runs out your instances are terminated automatically and their workspaces deleted, never silently overdrafted.
Stopping keeps everything on the node's NVMe, start it again and your files, models, and environment are exactly as you left them. Terminating releases the hardware and permanently erases local data, which is why it requires typing the instance name to confirm.
No. The hourly rate is the whole price, bandwidth in and out over the 200 Gb/s ConnectX-7 link is included.
Each instance is one whole physical DGX Spark, no virtualization, no shared GPUs, no noisy neighbors. SSH access is key-based only; password login for SSH is disabled across the fleet. Optional VNC desktop access, if you enable it at deploy time, uses a separately generated password for that connection only. Hardware is located in EU-Central.
Not yet, today every instance is a clean box you drive over SSH: install vLLM, SGLang, or whatever stack you prefer, and it all fits in unified memory. Preloaded model images that serve on boot are on the roadmap.
First top-up: pay $10, get $20 in credit — and refer a friend for another $50. Model preparation can take several minutes. $0 setup fee, no commitments, no egress charges.
Your GPUwerk Cloud account at a glance.
Manage your DGX Spark instances, start, stop, connect, terminate.
Choose hardware and an image, then review your access and running cost.
Commit to 3, 6, 12 or 24 months at a lower rate. Your first month is held on your card to secure it and charged once your dedicated Sparks are provisioned, within one business day. If you need more Sparks than we have racked, your first month is charged immediately instead, and we deliver within 14 days or refund it in full. You authorise your card once and each following month is charged automatically.
The last 200 lines your container wrote to stdout and stderr. Stopping shuts the container down and reserves this Spark for you at a discount off the hourly rate, so there is no new output until you start it again; nothing leaves the machine.
| Time | Event |
|---|
Credits, usage and transaction history.
Add prepaid credit now so it is ready when you deploy. Credit does not reserve a Spark or start an instance; live capacity is shown when you configure.
Demo mode, credits are added instantly without payment. First top-up: pay $10, get $20 in credit (we match up to $10) — and refer a friend for another $50.
When your balance drops below $25, is added automatically, billed at a 10% discount.
| Date | Description | Amount |
|---|
Share your link. When someone you refer makes their first qualifying top-up, you earn a matching credit too.
| Date | Status | Credit |
|---|
Authenticate against the GPUwerk API (api.gpuwerk.com/v1) to deploy and manage instances programmatically.
| Name | Key | Scopes | Created | Status |
|---|
Public keys installed on your instances at deploy time. Print yours with cat ~/.ssh/id_ed25519.pub, or create one with ssh-keygen -t ed25519.
Keys are saved to your account, so they follow you to another browser or machine. A key is written into an instance's authorized_keys when that instance is created and nothing rewrites it afterwards, so deleting a key here stops it being offered for future deploys, but whoever holds it can still log in to instances already running. Rebuild or terminate those to actually revoke access.
| Name | Fingerprint | Added |
|---|
Take a copy of everything we hold about your account, or close the account for good.
One JSON file: your account details, your full billing ledger and payment history, your instances and their event log, your SSH keys, and every message we have sent you. API key values are not included: we only ever store their hashes. Instance workspaces are not part of this export: a running or stopped instance's workspace lives on its node.
We're a small team running real hardware, you'll talk to a human.
We answer as quickly as we can during business hours. We don't publish a guaranteed response time or uptime figure; where one is agreed it lives in a separate written service level agreement.
Multi-node clusters, a reserved monthly node, and colocation of your own DGX Sparks are not self-service today. Email us and we will tell you what is possible and when.