Deploy AI models on H100, A100, H200, and AMD MI300X GPUs with up to 40% cost savings. Lightning-fast machine learning inference on enterprise GPU infrastructure.
14:07:02restoredcustom nodes, modelssame as you left them
14:08:20readyComfyUIopen in your browser
Podstop on one GPU, start on another
first session
RTX 4090 24GB
ComfyUI installed
next session
L40S 48GB
your setup restored
my-comfyuiready
running · ComfyUI is up on the first GPU
GPU machines that keep your work
01 / SYNC
Your working directory, straight to the GPU.
Push the directory onto the box, run it, see the output in your terminal. No image to build, no redeploy, no pipeline to wait on.
$ aq push
$ aq run -- python train.py
02 / PORTABLE
Stop here, start on another provider.
Stop a Pod in one datacenter and start it in another, on a different card. Files, models and packages come back the way you left them.
$ aq stop training-env
$ aq start training-env --gpu L40S
03 / SHAREABILITY
Share your environment with one link.
Send a teammate a link and they start a Pod with your exact packages, image and startup script. They get a fresh Volume of their own, and your files stay in your account unless you choose to include them.
$ aq env share training-env
AUTO-STOP · OPT-IN PER POD
The box you forgot about stops itself.
At or under 5% GPU for 20 minutes and you get a warning. At 30 the Pod's environment and volume are captured and the machine is released. Compute stops there; storage for the Pod keeps billing. You turn it on per Pod.
GPU 84% · ACTIVEGPU 3% · IDLE 20 MIN · WARNINGGPU 3% · IDLE 31 MIN · STOPPED
Observability on every deployment
Jobs and Pods alike. Utilization, VRAM, temperature and power, read off the card every few seconds. A reading the card cannot give you says so, instead of showing a zero.
