Configure and deploy a GPU instance with live marketplace pricing
Used in SSH prompts, dashboard, and billing. Letters, numbers, hyphens only.
We handle provisioning, networking, and monitoring. Best availability and fastest launch times.
Use your own Vast.ai API key. We orchestrate provisioning and monitor health. Lower cost if you have credits.
Use your own Lambda Cloud API key. Access reserved capacity if you have it.
Use your own RunPod API key. Access community and secure cloud tiers.
Use your own Hyperbolic API key. On-demand VMs and bare-metal clusters via their marketplace.
Use your CoreWeave K8s credentials. HPC-grade infrastructure with InfiniBand and VirtualServer provisioning.
Connect to a datacenter or colo you have direct access to. We handle orchestration over your network.
Connect to your own machine via SSH. Use for local GPUs, bare metal, or private infrastructure (e.g. on-prem training rigs).
PyTorch, Transformers, Accelerate, PEFT, bitsandbytes, Jupyter Lab, TensorBoard, wandb — ready for training and fine-tuning
JAX, Flax, Optax, Orbax, Jupyter Lab — for research and custom architectures
TensorFlow, Keras 3, TensorBoard, Jupyter Lab
vLLM server with OpenAI-compatible API — deploy and serve any HuggingFace model instantly
Stable Diffusion with ComfyUI node-based workflow, pre-loaded with popular models
Run and serve open-weight LLMs with a simple CLI. Great for inference and experimentation.
CUDA toolkit, cuDNN, NCCL only — bring your own framework and dependencies
Clean OS with only system packages. Full root access, install anything.
Zero-drop SSH via browser or native client. Session persists across disconnects.
Starts automatically on port 8888. Accessible via dashboard link — no port forwarding needed.
Full VS Code IDE in your browser via code-server on port 8080. Extensions, themes, everything.
Auto-generate public URLs for ports 7860-9000 (Gradio, Streamlit, FastAPI, etc).